Home

JUICE. To manually

1. Beals Smoothing t Relev No 166 237 0 072 S Picea obovata 1 A 0 068 Ptilidium pulcherrimum 3 0 065 Cladonia pleurota 9 0 047 Cladonia cenotea 31 0 042 Trientalis europaea 6 0 034 Abies sibirica 7 0 000 Vulpicidia pinastri 9 0 000 Lycopodium annotinun 61 0 000 Anemonoides altaica 6 0 000 Sphagnum wulfianum 9 Expected 0 506 Cladonia rangiferina 31 0 477 8 Betula rotundifolia 4 0 433 8 Cladonia arbuscula 31 0 412 Poa sibirica 6 0 402 Luzula sibirica 6 0 397 Polytrichum strictum 31 0 355 S Festuca ovina s lat 6 0 347 Cetraria islandica 31 0 300 Empetrum nigrum 61 us Average value 119 Export RED relev s Close Fig 78 Window with results of Beals smoothing for one relev The columns in the list are the Beals smoothing index the species name and the layer number The list is divided into a part with species actually occurring in the selected relev and a part with those that do not The Average value is the average of all indices of species occurring in the selected relev 2 6 Beta Diversity Traditionally species diversity is divided into three components alpha diversity which is defined as local diversity beta diversity as a between habitat diversity and gamma diversity also called regional diversity Beta diversity is a measure of difference between two or more local assemblages The v
2. 149 and 1324 Function InStr looks for matches in any part of the selected field It does not support wildcard characters For example to colour all relev s that mention Germany in their locality use Function InStr with Germany in the text box Note that this is equivalent to using Function LIKE with Germany in the text box But to select only relev s with localities that begin with the word Germany use Function LIKE with Germany in the text box 23 1 5 11 Searching Finding a certain species or relev can be difficult in larger tables The Find Species and Find Relev functions available in the Species or Relev menu or from the Icon Bar will display a text box in the Status Bar below the table Type in part of the species name and press the Find Species Head button The program will find the next species name that begins with that text and highlight the selected row Pressing the button again will find and highlight the next matching species Example The table sorted by layers contains the species Alnus glutinosa in three layers tree shrub and juvenile After the first selection of the Find button the program will display the part with the selected species in the tree layer the second selection will move the table to the species in the shrub layer the third selection will find Alnus glutinosa in the juvenile layer and the fou
3. Species response curves are widely used for description of species responses to an ecological gradient The response curve allows estimation of species optimum and also niche width tolerance identifying species as generalist or specialist Most of the widely used statistical methods assume that species response on a gradient has a symmetrical bell shaped Gaussian curve even though a number of studies have shown that this type of response occurs in real data quite rarely Several methods of dealing with the problem of modelling asymmetric species response curves were discussed in the study of Oksanen amp Minchin 2002a which was together with the detailed technical description in Oksanen amp Minchin 2002b taken as the basis for the algorithm built into JUICE and the R package software The R project macro for 91 modelling species response is still being developed A more detailed and current description of the procedure is available at http botanika bf jcu cz david hof php 2 8 1 General Information about the Function There are various methods of modelling species response curves Those implemented in JUICE are discussed here 2 8 1 1 Models of Species Response Curves Available methods for modelling response curves in JUICE include 1 Bell shaped Gaussian curve with traditional symmetrical shape 2 Generalized linear models GLM with polynomial of 1st 2nd or 3rd degree 3 Generalized additive models GAM with
4. button All selected fields appearing in the right list box will be imported 1 4 2 Cornell Condensed CC Format The Cornell condensed format file consists of three parts table data species abbreviations and relev numbers Several types of CC files exist they differ especially in the table data format defined in the second line of the file JUICE should accept any of these formats as long as they conform to the following convention the first number of the line gives the relative relev number and the remainder of the line consists of pairs in which the first number refers to the species and the second number to the species cover Cover values may be expressed as percentages or as categories on an ordinal scale 1 9 The species abbreviation section contains ten abbreviations per line Each abbreviation consists of either 8 characters or 7 characters plus 1 character for the layer number The third section of the file is reserved for relev identification numbers Each number has 8 characters and there are 10 numbers per line A sample CC file can be downloaded from the JUICE web page and tested directly in the program Example Juice analysis T I5 5 I5 F8 1 5 1 1 2 0 2 2 0 S 13 0 4 1 0 5 38 0 1 6 2 0 7 2 0 8 3 0 9 1 0 10 1 0 1 11 3 0 12 3 0 13 3 0 14 3 0 T5 ESO 1 16 2 0 17 1 0 18 3 0 19 13 0 20 2 0 1 21 2 0 22 1 0 23 1 0 24 2 0 25 2 0 1 26 2 0 27 2 0 28 2 0 29 2 0 30 13 0 2 1 1 0 5 2 0 10 2 0 16 3 0 21 2 0 2
5. d standardization of the size of all C C site groups target ee group being c B ofthe sare size ham mamma poama as the others m poa 8856578 eee Fig 42 A scheme of different standardizations of the size of relev groups Each line represents a data set segments are groups labelled A B C and D and segment length corresponds to the number of relev s in each group The thick part of each segment represents relev s that contain the given species the thin part represents relev s where the species is absent In figures b to d the four lines represent the four standardizations used for calculation of species fidelity to the target groups A B C and D Part a of the figure shows a simple artificial data set with vegetation units of unequal size The phi coefficient depends on the size of the vegetation units Standardizing the data set removes this dependence After the standardizations in Fig 42c or Fig 42d the fidelity measure depends only on the relative frequencies within each relev group The standardized size of the target vegetation unit can differ from the size of the other vegetation units Fig 42c or be the same as the other vegetation units Fig 42d The phi values after such standardization are entirely independent of the size of the vegetation units and can be directly compared across different vegetation units In some cases however it may not be desirable to standardize the size of all
6. Cluster Analysis Setup m will start PC ORD installed in Relev s Used in Analysis omp I cl Data Transformation value 0 100 Minimum 2 Maximum 1 Distance Measure Group Linkage Method Sorensen Bay Curtis Nearest Neighbor Relative Sorensen Farthest Neighbor Jaccard bd Median E L Group Average b Euclidean Pythagorean oniro g Relative Euclidean Ward s Method di Corellation Ld Flexible Beta 0 25 Chi squared McQuitty s Method Path of the output fi C Program Files TAX 2000 pcord csv Cancel Continue gt gt gt Fig 64 Window for managing PC ORD cluster analysis The user selects relev s to be analysed by PC ORD by choosing a colour The Data transformation options are usually used to reduce the weight of higher cover values Pseudospecies cut levels can be used to convert from percentage cover values to an ordinal scale with a small number of categories In the example above 0 4 would be converted to 1 5 24 would be converted to 2 and values of 25 and above would be converted to 3 The Floating cut level by Species data value 0 100 option converts a cover value to 1 if it is less than the value in the Species Data Column or 2 if it is equal or greater This can be used e g when each species s median cover has been stored in the Species Data Column to analyze clusters based only on whether species cover is below the median or not Such a standardisat
7. Sorting menu select Sort relev s by clusters and SYN TAX 2000 8 Open files DENDROGRAM DEN and SEQUENCE DAT It is then possible to proceed as described in Section 2 2 1 3 2 3 COCKTAIL Method Milan Chytry amp Lubomir Tichy The Cocktail method Bruelheide 1995 2000 was designed to simulate the Braun Blanquet approach to classification in which classification is based on expert knowledge and not on unsupervised algorithms such as those used in e g TWINSPAN or cluster analysis computer programs With the Cocktail method as implemented in JUICE an expert makes subjective choices during the classification process while the program suggests possible solutions and ensures that particular steps in the classification process are applied consistently throughout the data set Delimitations of the resulting vegetation units are explicitly formally described which means that new relev s that were not present in the original data set can be unequivocally classified as belonging or not belonging to a particular vegetation unit 75 The Cocktail procedure starts by defining groups of species that tend to occur together in relev s of a large database Using a large database that covers a broad spectrum of different habitats and a large geographical area is important for obtaining species groups of more general validity Species of the same group usually have similar habitat requirements and phytogeographical affinities Cocktail
8. Synoptic Table menu select Sort Species In Synoptic Tables and press the button next to Zlatnik s value 1 11 8 Average Cover Barkman s Total Cover Value The average cover AC within a relev group as defined by Barkman 1989 is AC SUM Cover number of relev s Note The number of relev s includes those with a cover value of zero i e where the species is absent JUICE c documents and settings lubomir tichy dokumenty 0_lubos juice kurz 10_3_2005 dyje_valley2 wct File Edit Species Synoptic Table Table Simulation Help black v 4 E IX Phi cost Total time 3 days 16 h 1 min 45 sec Number of releves releves 202 Species 631 Ulmus glabra Galium aparine Phalaris arundinacea Lamium maculatum Carex remota Lunaria rediviva Fagus sylvatica Fraxinus excelsior Carpinus betulus Carex sylvatica Brachythecium rutabu Tilia platyphyllos Dryopteris filix mas Acer platanoides Galeobdolon montanum Polytrichum formosum Melica uniflora Carex pilosa Convallaria majalis Stellaria holostea Avenella flexuosa N OOMNMOOO DOWWWW H OFPONWWHEEBHOO oono o oo NHO OOoooooo Frequency 8 Rel flo Row Turboveg No Column 19 Ulmus glabra Fig 53 Synoptic table with average covers 1 11 9 Combined Synoptic Tables Standard synoptic tables contain only one important aspect of the data set while others are not presented Information such
9. scceseseseeseceseesecseecseeeseeeeeeseceeenseeesecesecnsecneecneeeeeeneeass 91 2 8 1 2 Technical Notes on Particular Modelling Strategies sss 91 2 8 1 3 How species optimum and tolerance are calculated 0 0 ccceeceeseeeeceseceeceeceeenseeeseeeeeeeeneeees 92 2 8 2 Installation of the Function JUICE ett aoe Bee 93 2 8 3 Calculation of Species Response Curves SRC in JUICE sess 93 LITERATUR Bis vccsectcccccccccccccenweceneuncencwecnveneweedveueseentoueceuvweunsvenweueceieweueseunneueweurweuenveuwens 96 iii Acknowledgements Acknowledgements We thank Milan Chytry as the first tester of new program versions and designer of many functions Many thanks to Zoltan Botta Dukat Heike Culmsee Michal Hajek Petra Hajkova Rense Haveman Marcela Havlova Stephan M Hennekens Eszter Illyes Florian Jansen Ilona Knollov Martin Ko Petr Pet k Honza Role ek Urban Silc Stephen S Talbot David Zeleny VaSek Zouhar and all others who helped to make the program and this manual better This study was funded from the grants GACR 206 99 1523 GACR 206 02 0957 GACR 206 05 0020 MSM 143100010 and MSM 0021622416 iv 1 Getting Started 1 1 Introduction JUICE is a Microsoft WINDOWS application for editing classifying and analysing large phytosociological tables It includes many functions for easy manipulation of table and header data The program is optimised for use with TURBOVEG software Hennek
10. Mefford 1999 This option is available only when a Percentage Constancy synoptic table is displayed 43 Option 1 opens the following window where the parameters of exporting the table may be modified Export species in original colours ll Highlight important frequency or fidelity values Bold style of important frequency or fidelity values Cancel Continue gt gt gt Fig 38 Window managing synoptic table export Synoptic tables may be saved in single or combined form See Section 1 11 9 All data will be saved similarly to the way they are displayed on the screen 1 9 5 Other Exports into the RTF Export File The current RTF export file is opened for running exports of interspecific associations created during COCKTAIL classification and Matching More information about COCKTAIL classification is written in the Chapter 2 3 1 9 6 Special Export Formats The Export submenu of the File menu includes the option to export the table in a variety of other special formats enabling the data processed by JUICE to be analysed more precisely The program supports table export as a Cornell condensed file a MULVA input file SYN TAX files or a MATLAB file A text export of similarity indices of relev s to constancy columns is available from the menu Analysis and Matching See the Chapter 2 4 2 3 1 9 7 Export for D MAP File D MAP Morton 2005 is simple GIS software which is very useful fo
11. New similarity indices for the assignment of relev s to the vegetation units of an existing phytosociological classification Plant Ecology 179 67 72 Westhoff V amp van der Maarel E 1980 The Braun Blanquet approach In Whittaker R H ed Classification of plant communities Dr W Junk Hague 98 Wildi O amp Orl ci L 1996 Numerical exploration of community patterns A guide to the use of MULVA 5 2 edition SPB Academic Publishing bv Amsterdam Zar J H 1999 Biostatistical analysis 4 edition Prentice amp Hall Upper Saddle River
12. See Section 1 6 1 2 Classify the table and display it in the form of a synoptic table using either frequency or fidelity 3 From the Synoptic Table menu select Compare Two Synoptic Tables and Save Image Of Synoptic Table to save the table 4 Load the second table and display it in the same form as the first one 5 From the Synoptic Table menu select Compare Two Synoptic Tables and Load Image And Compare The program will display the resulting distance A crossing table of all Euclidean distances is saved onto the clipboard 68 2 Data Analysis 2 1 TWINSPAN TWINSPAN Two Way Indicator Species Analysis is a numerical classification method developed specifically for hierarchical classification of community data The technique is based on the concept that a group of relev s will have a corresponding group of indicator species that characterize that type Species and relev s are sorted based on a reciprocal averaging algorithm Hill 1979 TWINSPAN is included in the standard JUICE installation package and is installed in the JUICE application directory This software runs as a stand alone program in a simulated DOS environment Numerical classification by TWINSPAN can be accomplished either 1 directly from JUICE or 2 through export of a CC file manual classification in TWINSPAN and import of the results 2 1 1 TWINSPAN in JUICE Before the function is started it is necessary to giv
13. a relatively high phi value can be attained even with a smaller difference in frequency between the target and other units 0 0 2 0 4 0 6 0 8 n n N N 0 0 2 0 4 0 6 0 8 n nJN N Fig 40 Dependence of the phi coefficient on the relative frequency of species occurrences within vertical axis and outside horizontal axis the target vegetation unit shown for vegetation units equal to a 1096 and b 5096 of the size of the total data set c The Dufr ne Legendre Indicator Value Dufr ne amp Legendre 1997 is defined as the product of relative species abundance and relative species frequency within the target vegetation unit Relative species abundance is the mean abundance of the species in the target vegetation unit compared to all vegetation units in the data set the sum of mean abundances in each vegetation unit is used instead of the sum of actual abundances over all relev s in order to remove the effect of unequal vegetation unit size JUICE only computes the Dufr ne Legendre Indicator Value for presence absence data which means that the relative abundance is replaced 47 by the relative frequency in the target vegetation unit divided by the sum of relative frequencies in all vegetation units of the data set For a comparison between a target vegetation unit and the rest of the data set the Dufr ne Legendre Indicator Value can be expressed as n N ny IndVal n N n n
14. gt JUICE c juice_fin 1 louky_ass_str_del01 wct File Edit Species Relev s Head Sorting Separators Synoptic Table Indicator Values Analysis Table Simulation Help Edge fe RO EE i EAE x a ing Statistics u value hyp jad 77 5 E a Total time 2 days 0 h 38 min 34 sec Running number 5 the Shert headers Releves 1210 688 5964999 9 1974111121112222 Species 606 765434639210087139933236134965567207283206576237864123415690123 Achillea aspleniifol S Achillea millefo 144144112141 r 1 1 22 2 222 21 11 21 1 22 1 11 1 122 2 S Anthoxanthum odo 1 1 31 132 r13 1 1 2 122 22 221 224 2 1221 121211223122 Alopecurus pratensis 1 ee 11111 1 11 1 11 Achillea ptarmica g CoS M EPA MEE EE DU EE E DEDE Acinos arvensis ebd ercp od EM D EIN E d Ce m S Aconitum napelus MW LIGNE SUELE EUER m EDI Ic eee ee ey ara S Adoxa moschatellina TUUS DUC STER STE coe Gor PESCE B A WEN RC A RIN AS S EMITS ARC OOS Poy tata Ty ERES TUO HE ERO T ee A ri w Ecl c ace px uM i to z gb ers Aqromyron caninum Ap M ee P RMR ees pahe nina 0O apne died A i ris Ursus us t Aarostis Ss o Meere d eR accts petiole ROME OI CRUDO aod e JL CUu EON Ii Ajuqa qenevensis j WH DEBEMUS EI M CLE E secte ume tees el ial ae aisles Ajuqa reptans KL ecd cha pitta sh aa AL m NT A PINO ee ES se S Alchemilla hybri I ROL RR TN ACE eR D EUTROPIO CP Ede t dre edd tee lur SEE S Alchemi
15. ssessssessesee eee enne 77 2 3 4 Cocktail Groups Definition ssiri inrcis iieii iieii dete ie idees 78 2 3 5 Cocktail Algorithm for the Definition of Species Groups sse 80 2 3 6 Definition of Relev Units by the Combination of Cocktail Groups sees 81 2 3 7 Exp rt Syste rcn teii M 82 2 4 Similarity Indices 84 2 4 1 Desctiption iuis nei eee eite ee 85 2 4 2 The use of similarity indices in JUICE Matching function 86 2 4 2 1 Procedure description ee deett ei ie ete de ehe ade 86 2 4 2 2 Relev sort by similarity indices sess 86 2 4 2 3 Export of similarity index values sess 87 2 5 hITIuunnurn sesiis 87 2 5 1 The Calculation of Beals Smoothing in the Program sssssssseseeeeeeeeeennns 88 2 6 hirgpiuoda 88 2 6 1 How to Measure Beta Diversity in JUICE essere ener 88 2 7 lo ninugpdre c 89 2 7 1 How Euclidean Distance Is Calculated eese 89 2 8 Species Response Ela e M M 90 2 8 1 General Information about the Function essent nennen 91 2 8 1 1 Models of Species Response Curves
16. 0 9 Each dot is a species frequency inside and outside the selected first vegetation unit The user can select parameters and graphically test each vegetation unit 1 10 6 Quantitative Fidelity Measures JUICE versions 6 3 66 and higher can calculate fidelities not only from presence absence data but also from quantitative cover values For quantitative calculation variables are defined as n sum of all covers of the species in the data set n Sum of all covers of the species in the target vegetation unit Covers range in the interval 0 1 The fidelity measure calculations are analogous to those described above in Section 1 10 1 The choice between presence absence and quantitative fidelity measurement is available from the Options window See Fig 44 Note Quantitative fidelity measures are only available for synoptic tables Functions COCKTAIL Groups and Interspecific Associations are calculated using presence absence data 54 1 11 Synoptic Tables Synoptic tables summarise the results of any classification of relev s Therefore they are one of the basic results of all phytosociological analyses and studies They give an overview of classified vegetation units in the data set and help the researcher understand sophisticated relations among species in context with environment Synoptic tables can contain several types of information which can be used for additional sorting of species and analysis of d
17. 1979 TWINSPAN a FORTRAN program for arranging multivariate data in an ordered two way table by classification of the individuals and attributes Ecology and Systematics Cornell University Ithaca New York Juh sz Nagy P 1964 Some theoretical models of cenological fidelity I Acta Biol Debrec 3 33 43 97 Knollov I Chytry M Tichy L amp Hajek O 2005 Stratified resampling of vegetation plot databases as a bias reduction in classification studies J Veg Sci 16 479 486 Koleff P Gaston K J amp Lennon J J 2003 Measuring beta diversity for presence absence data Journal of Animal Ecology 72 367 382 Kubat K ed 2002 Kli ke kv ten Cesk republiky Academia Praha McCune B 1994 Improving community analysis with the Beals smoothing function Ecoscience 1 82 86 McCune B amp Keon D 2002 Equations for potential annual direct incident radiation and heat load J Veg Sci 13 603 606 McCune B amp Mefford M J 1999 PC ORD Multivariate Analysis of Ecological Data Version 4 0 MjM Software Design Gleneden Beach Oregon 237 pp Morton A 2005 DMAP for Windows Version 7 0 Alan Morton Winkfield Windsor Berkshire Computer program http www dmap co uk welcome htm Oksanen J amp Minchin P R 2002a Continuum theory revisited what shape are species responses along ecological gradients Ecological Modelling 157 119 129 Oksanen J amp Minchin P R 2002b Non linear
18. 2 blue 3 sea green 4 green 5 yellow 6 violet 7 grey 1 7 4 Transformation of Species Data Numerical values n in the Species Data Column can be replaced by their squares n square roots Vn or multiplicative inverses 1 n 1 7 5 Statistics Summarizing Relev Data Short headers can contain data about the relev s such as means of Ellenberg indicator values sums of species statistics Shannon Wiener indices potential heat load etc See Section 1 8 A statistic summarizing the short header data from the relev s in which the species occurs such as minimum maximum mean or median value can be written to the Species Data Column From the Species menu select Species Data and Short Header Data This gives a menu of summarizing statistics to choose from 1 7 6 Ellenberg Indicator Values Ellenberg indicator values can be written to the Species Data Column from the Species menu select Species Data and Indicator Value This opens a menu for selecting whether to display indicator values for Light Moisture Continentality Temperature pH or Nutrients Before using this function it is necessary that the indicator values be defined More information on Ellenberg indicator values will be included in later editions of this manual 1 7 7 External Species Data Other species data can be imported into JUICE if they are in a suitable for
19. 2 tree layer middle 3 tree layer low 4 shrub layer high 5 shrub layer low 6 herb layer high 7 juveniles 8 seedlings 9 moss layer Layer is the most important information about the species It is displayed automatically at the bottom Status Bar near the selected species s name Layer can be displayed in the Species Data Column from the Species menu select Species Data and Layer View or select the Layer Icon on the Icon Bar See Section 1 5 1 Layer can be displayed as a number as text or both 1 7 2 Frequency and Cover Values The Species Data submenu from the Species menu has functions for writing Frequency Maximum Cover or Median Cover to the Species Data Column Frequency is the frequency of species occurrences in the data set Maximum Cover is the species s maximum cover value in the data set Median Cover is calculated from all non zero cover values These statistics can also be viewed by selecting Species Statistics from the Species menu To return to Standard Display it is necessary to select Species Statistics again 1 7 3 Sequence and Species Colour The Sequence function writes consecutive numbers into the Species Data Column The No of Species Colour function writes each species s current colour to the Species Data Column according to the following code 0 black 1 red
20. 2 and so on Thus relev s with consecutive indices are not necessarily in adjacent grid boxes on the virtual map 1 8 2 6 Potential Annual Direct Irradiation PADI And Heat Load The function Potential Annual Direct Irradiation And Heat Load can be used to calculate PADI or heat load for relev s based on slope aspect and latitude as proposed by McCune and Keon 2002 37 There are three available models Model 1 In Rad MJ cm yr 1 467 1 582 cos lat cos slo 1 500 cos asp sin slo sin lat 0 262 sin lat sin slo 0 607 sin asp sin slo Model 2 In Rad MJ cm yr 1236 1 350 cos lat cos slo 1 376 cos asp sin slo sin lat 0 331 sin lat sin slo 0 375 sin asp sin slo Model 3 Rad MJ cm yr 0 339 0 808 cos lat cos slo 0 196 sin lat sin slo 0 482 cos asp sin slo where asp is aspect slo is slope and lat is latitude Potential Annual Direct Incient Radiation And Heat Load direct incient radiation Models Model 1 Model 2 Model 3 Test of the method Latitude example 485020 7 48 50 20 PADI Radiation 0 137 In RAD Slope Heat Load 1 148 In HEAT Aspect pect Recalculation Saved values e Radiation Heat load Add to short headers Fig 33 Calculation of potential direct incident radiation and heat load The window allows the user
21. 23 3 0 25 2 0 26 13 0 31 3 0 32 3 0 2 33 2 0 34 2 0 35 13 0 36 1 0 37 13 0 2 38 2 0 39 3 0 40 3 0 41 13 0 42 2 0 2 43 2 0 44 1 0 3 4 3 0 5 3 0 6 2 0 10 2 0 13 2 0 3 16 3 0 18 2 0 21 2 0 23 2 0 25 3 0 3 26 2 0 31 3 0 33 2 0 35 13 0 36 2 0 3 38 3 0 39 3 0 40 2 0 41 38 0 45 2 0 3 46 1 0 47 2 0 48 1 0 49 2 0 50 1 0 3 51 2 0 52 13 0 53 2 0 54 2 0 4 2 1 0 3 2 0 4 2 0 5 38 0 6 2 0 4 10 3 0 13 3 0 20 1 0 23 13 0 25 3 0 4 26 2 0 27 2 0 30 38 0 36 2 0 42 3 0 4 55 2 0 56 2 0 57 3 0 58 2 0 59 2 0 4 60 3 0 61 2 0 62 2 0 63 1 0 64 2 0 4 65 1 0 66 2 0 0000 ACHI MI6AGRIEUP 6ARRHELA6ASTEAME 6BRAYP IN6BUPLFAL6CENTSCA6CLINVUL6CORUSAN7CRATMON7 DACYGLO6ELYMREP 6FESTRUI 6FRAGVES 6GENITIN6HIERP IL6HIERSAB6KNAUKIT6MEDIFAL60RIGVUL6 PIMP SA6PLAALAN6PRUUFRU6RANUPOA6SALVPRA6SANGMIN6STACREC 6TARASEO6VERBC A6VICITEN6 ALYSALY 6AREN SE6ASPRCYN6CHAARAT 6FESTVAL6FRAGVIR6GERASAN6 INULENS 6KOELMAC 6MEDIMIN6 POTEINC6ROSA CN7 TEUCCHA6THLSPER6ARENSER6COLTARB 7MEDI VA6PRUUSP I 7ROSAP IM6RUBUCAE6 SCABOCH6SESEOSS 6STAC RE6THYU PA6ACERCAM7ASTAGLY 6CAREMIC6CARIBET4CRATMON4CYTINIG6 ERYNCAM6EUPHPOL6GALUVEU6HELINUM6P ICRHIE6ROSARUB4 400179 400180 400181 400182 Header data are imported automatically if header data files have the same name as the CC file and have a correct format See Section 1 4 6 The import process starts with selection of the CC file Then the following window appears El Table file The selected file is an output from TURBO other file in Cornell Condensed f
22. AJ o te ak E GD lt Ctrl gt x HHA p usa fs ix Statistics Phi coeff A Total time 1 day 0 h 13 min 19 sec Option Button Relev colours for fast selection Species colours for fast selection Note 1 The buttons Reset relev colour and Reset species colour have two functions One click of the left mouse button will reset the currently selected colour while double clicking will reset all colours See Sections 1 5 2 and 1 5 3 for more information on colours 11 Note 2 The Option Button opens the Options window also available from the File menu Its purpose is to provide quick access to the Fidelity tab see Section 1 10 3 but other option tabs are available such as the Display parameters tab described in Section 1 5 8 below The Status Bar at the bottom of the window contains information about the last species selected its order in the list its full species name to a maximum of 50 characters its layer number and its total frequency in the data set The values Rel No the relev s relative number in the imported data set and Turboveg No a unique TURBOVEG number in the range 1 999 999 refer to the most recently selected relev The Row and Column values give the current position of the cursor in the table Frequency 440 Rel No Turboveg No 11 Agrostis capillaris 6 1 5 2 Mouse Keyboard Functions Working with table data requires the use of
23. Agrostis capillaris Agrostis vinealis Ajuga genevensis Anthemis ruthenica Anthoxanthum odoratum Anthyllis vulneraria Apera spica venti Arenaria serpyllifolia agg Armeria vulgaris subsp vulgaris Arrhenatherum elatius me e This function repeats automatically selected merging deleting and undeleting steps used at previously mana ged tables Achillea millefolium agg Achillea collina Achillea millefolium Achillea pannonica Achillea pratensis Achillea millefolium sub Achillea millefolium sul Achillea pannonica Achillea millefolium Achillea collina Achillea millefolium agd CQ O0 0 0 0 0 0 0 0 A AE Fig 28 Autorepeat function The Autorepeat Function window will then contain two lists On the left is a list of the steps that were performed to edit the selected file On the right is a list of the species that were affected by the highlighted step Species prefaced by were present in the model file but are not present in the current file Pressing the Run Selected button will cause all the editing steps from the model file to be performed on the current file Alternatively the user can choose step by step to either perform the highlighted step Run Step or skip it Omit Step Note This function is most useful when the current table has exactly the same species as the previously edited table When the current table has additional species the results will often be unexpec
24. Bar 2 Open the Head menu and select Relev Colour According To The Head This will open the following window Relev Colour According To Header Data Selected colour Select header field Date year month day Syntaxon code Relev area m2 Altitude m Aspect degrees Slope degrees Cover tree layer Cover shrub layer Cover herb layer Cover moss layer Locality Longitude Write the value stands for character string 1992 Function LIKE Function InStr Selected relev s 133 ETE Fig 19 Window for colouring relev s by header data 3 From the list select the relevant field 4 Inthe text box type the text to be matched 5 Hit the Continue button Relev s matching the text are given the indicated colour The number in the Selected relev s field indicates how many relev s match the text JUICE provides two text matching functions Function LIKE requires exact matching of the text in the box but it supports wildcard characters A symbol represents any character and a symbol represents any string of characters For example 1 can be used to select all relev s at an altitude of 100 199 m because it matches text like 132 and 149 but not 711 the first character is not 1 nor 1324 the text is more than 3 characters long The text 1 would match all numbers beginning with 1 including 1 13
25. Fig 4 Options Check List Import The button Open new check list file allows the user to specify a new species list file The ID Number Abbreviation and Species Name boxes are used to define the number of characters reserved for each field If the first line of the file defines the length of these fields the values appear in the boxes automatically The file format can be tested by clicking on the Test file structure button Check list encoding must be switched on for the NEWFLORA TXT file Other files are not encoded 1 4 4 Spreadsheet Format A spreadsheet format file contains the title of the table first line the number of relev s the relev numbers and the table itself species name layer and cover codes Cover codes may be characters from the Braun Blanquet scale or any other or percentage values The Import Manager begins by prompting you to select a file When the file is selected you are informed about the title of the table the character used to delimit columns layer information and table size steps 2 through 5 If the file contains only species names without layer information uncheck the box in step 4 In step 5 make sure that the indicated table size is correct In step 6 make sure the table corresponds with real data The last step is to specify the cover values If the scale is in Braun Blanquet codes or percentage values simply select the appropriate option Otherwise cover v
26. Gi NE de ND QU GU GO GI e de GUNT IO GENE E UID AKHWUAOOCOMACOMHEUNUSSUEEOURONE THOS L 8 a 8 8 8 8 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 9 D 9 9 9 9 9 9 auohavapunuoaacoaanvahooaanadanana oN ocouoocuD vaobDouoocoooooono OhNNOHeNWOSNOVeCONOUSUEROCD HPHROOUPPHHeNZUOD NMueercoOOsROugewcuow Fig 76 Similarity Index Display mode The function can be illustrated with a small example The data set in Fig 76 is divided into two parts columns 1 through 8 were classified by TWINSPAN while column 9 is the group of new relev s to be analysed by the Frequency Positive Fidelity index The column number with the maximum value and the quality of this information are displayed in the two rightmost columns Max and Qual The quality index is the sum of all similarity indices for the relev divided by the maximum value The index has a minimum value of 1 and a maximum value equal to the number of columns A value near 1 indicates that the relev fits well only with one column while values near the maximum indicate that the relev fits equally well or poorly with each column 2 4 2 2 Relev sort by similarity indices In Similarity Index Display mode the program can sort relev s by similarity indices From the Analysis menu selecting Matching and then Sort Relev s opens the window shown in Fig 71 The user can sort all relev s of the designated colour according to t
27. Headers with Header Data essen nennen nennen nennen nnns 35 1 8 2 4 Shannon Wiener Index and Evenness essent eene 35 1 8 2 5 Geographical Position Index sessssssssssesseseeeeeeenen eene enne ener nnns 36 1 8 2 6 Potential Annual Direct Irradiation PADI And Heat Load sees 36 1 8 2 7 Sum Average Minimum Maximum and Multiplication of Species Data 37 1 8 2 8 Ellenberg Indicator Values ccccecsceeseeesceescesecesecesecseecaeeeseeeneceneeeeeseeneeensecesecaeenueceecneeeneeees 38 1 8 2 9 Importing External Short Header Data sse eren ener 38 1 8 3 Colouring Relev s According to Short Header essere 39 1 8 4 Short Header Averages Minima and Maxima sese 39 1 8 5 Short Header Sorting 4 cce ies e dec HER e Re He RE eee dee 40 1 9 I qrnlrdbripcpEe REM c 40 1 9 1 Saving Files in JUICE teet iti eere tic RH dee ife ed dee dde 40 1 9 2 Th Current Export File eee e eee it e dete He ee RD 41 1 9 3 Table BXpott niei hie eater e ree ee d PE RE ERR 41 1 9 4 Synoptic Table Export ecd it duces eee ce Te aee Fee ddnde ae 42 1 9 5 Other Exports into the RTF Export File 43 1 9 6 Special Export Form ts nennen neii te e eerie de iet i eiae dene deiode 43 1 9 7 Export for D MAP File eite eee Et e i eee e ede ce 43 1 9 8 Species D t Export ecce eee eee eite estes eee diee
28. Save and enter a name for the file Actually three files are generated when a table is saved a WCT an STR and an EXP file The WCT file has a special binary structure not suitable for manual editing but the other two are simple text files containing header data in the same form as they were imported See Section 1 4 6 Note Although the File menu does not explicitly have a Save As option Save actually works the way Save As does in many other programs A dialog box will open and you will have the opportunity to either confirm that you want to save the file under its current name or to enter a new name Thus it is possible to save an altered file under a different name without overwriting an older version of the file Furthermore before JUICE overwrites an existing file it asks for confirmation 41 1 9 2 The Current Export File JUICE s standard export is a rich text format RTF file to which subsequent exports are appended without overwriting previous exports This is useful for exporting running results from interspecific associations or COCKTAIL analysis for example The file into which JUICE will export appears at the top of the Export menu available from the File menu Clicking on this Current File function opens a window where the current export file name can be changed If the name is not changed but the Save button is pressed JUICE interprets this as a command to wipe the export fil
29. Synoptic tables Display parameters Combined synoptic tables Select one from seven 45 3 ED 1 combinations of frequency Fei i Fiddity C fidelity and cover idelity s k Frequency Category Fidelity Fidelity Freq Frequency Frequency Category Fidelity A ED EDL MR Cae Lower Higher Thresholds Fidelity threshold 40 50 Important fidelity or frequency values in v Display colours Colour f Colour synoptic tables are highlighted by selected background colour Frequency threshold 70 60 lv Display colours Colour J Colour Fig 47 The Synoptic tables tab of the Options window 55 For more sophisticated synoptic table graphic display two thresholds with different colours can be defined Values greater than the Lower threshold will be given one colour and values which exceed the Higher threshold will be given the other colour For single colour highlighting the Lower threshold can be set to a value equal to or greater than the Higher threshold Colours can be completely switched off with the Display colours check box Note If a value is entered that causes the Lower threshold to be greater than the Higher threshold the Higher threshold is changed automatically This means that it may be impossible to lower the Higher threshold without first lowering the Lower threshold 1 11 2 Percentage Constancy Synopti
30. and 3 classify this part of the table once more If the Species sorting box is checked species will be sorted by reciprocal averaging If the Make separators box is checked hierarchical separators will be included in the table See Section 1 5 4 Instead of the Species sorting option we recommend sorting using the Sort Species in Synoptic Table function See Section 1 11 11 2 12 Use of TWINSPAN as a Stand Alone Program Some users wish to use more sophisticated classification features of TWINSPAN e g to omit certain species or to see eigenvalues species or relev indicators and other information about each division In such a case it is possible to run the TWINSPAN program manually on a JUICE table exported in Cornell Condensed file format See Section 1 9 3 The TWINSPAN output files with suffixes TWI and PUN are useful for additional sorting which is available from the Sorting menu select Sort the Table by Clusters and TWINSPAN 2 2 Cluster Analysis Three widely distributed programs for multivariate data analysis are integrated with JUICE for cluster analysis PC ORD McCune amp Mefford 1999 MULVA Wildi amp Orl ci 1996 and SYN TAX 2000 Podani 2001 2 2 1 Cluster Analysis via PC ORD PC ORD is a Windows program that performs multivariate analysis of ecological data In addition to utilities for transforming data and managing files PC ORD offers many ordination and c
31. are usually less useful for table explanation Therefore the program displays all values of zero or less as dashes 1 11 6 Synoptic Tables and Cover Maximum Average Median Modus A synoptic table can also be used for presenting cover values These are selected from the Abundance Value submenu of the Synoptic Table menu As an example suppose a species has the following cover values in a group of 11 relev s percentage number with cover code 1r 16 160 2 3 42m 8Qa 8Qa 18 2b 38 3 63 4 The maximum cover for this species will be 63 the average will be 13 rounded the median will be 4 and the modus the most frequent value will be 1 These statistics only consider relev s in which the species is present To find the average cover over all relev s in the group use Barkman s Total Cover Barkman 1989 as described in Section 1 11 8 below Note Cover values in a synoptic table are always displayed as percentage numbers 1 11 7 Zlatnik s Combined Scale Another type of table available from the Abundance Value submenu is Zlatnik s Combined Value Pl va amp Pr a 1969 Czech foresters use a special synoptic table in which values are defined according to this table 58 Zlatnik s Combined Scale Display this form again Fig 52 Information window with Zlatnik s scale These values combine species relative frequencies with maximum cover To see this table in JUICE open the
32. as frequency fidelity and cover can be presented in multiple tables 59 but this takes up space and it can be difficult to see relationships between the tables JUICE can display two values in a combined synoptic table thus overcoming these disadvantages Selecting Combined Synoptic Table from the Synoptic Table menu displays a combined General External program paths Check List Import Fidelity measures Synoptic tables Display parameters Combined synoptic tables Select one from seven ED combinations of frequency 6T E Cav Fidelity C GL fidelity and cover wen Frequency Category Fidelity Ge EDS Ie Er Frequency Frequency Category Fidelity Lower Higher Thresholds Fidelity threshold 40 50 Important fidelity or frequency values in v Display colours Colour Colour synoptic tables are highlighted by selected background colour Frequency threshold 70 60 IV Display colours Colour Colo Fig 54 Selecting a combined synoptic table from the Synoptic tables tab of the Options window From this tab one of seven combinations of frequency fidelity and cover can be displayed The highlighting thresholds for frequency and fidelity see Section 1 11 1 are applied to their own columns in the table For example if the frequency and fidelity combination is selected the combined table will look like this JUICE c documents and settings lubomir tichy dokumenty 0_lu
33. check box is checked Dominant species are defined as all species that have cover values higher than the cover threshold A species appears in this list if its cover exceeds the threshold value in any relev of the selected vegetation unit The list of dominant species can be limited to 62 species that exceed a certain minimum relative frequency as specified in the Minimum freq box Species are displayed with layer number and fidelity or frequency value If the Show references check box is checked species that appear in more than one list will be cross indexed to the other list or lists Dg for diagnostic C for constant and Dm for dominant Note 1 The display is not updated when a display mode check box or a threshold value is changed To enact the desired display changes press the Refresh button Note 2 Changing the Fidelity threshold or the Frequency threshold changes the threshold throughout the program just as though it were changed in the Synoptic tables tab of the Options window Species in the lists can be selected with the mouse or keyboard and assigned a selected colour in the table by pressing the Mark in the table button Note Shift click can be used to select a block of species In particular to select an entire list click on the top species and then Shift click on the bottom species Ctrl click can be used to select multiple species which are not
34. diagnostic constant species Positive Fidelity index PFDI is a measure based on the species fidelity concept proposed by Chytry et al 2002 The index is calculated in the same way as the Frequency Index FD is the positive fidelity value phi coefficient between a particular vegetation unit and a species i present in the analysed relev Chytry et al 2002 All negative fidelities are neglected pri 100 gt Fo Jr where FD gt 0 Eq 16 ieRAC ieC The Positive Fidelity index 1s defined in the interval 0 100 The index value depends on the number and diagnostic information degree of fidelity of the diagnostic species thus overcoming the disadvantage of the Frequency Index but it discriminates poorly between relev s composed of widely distibuted constant species that are also found in the target vegetation unit and relev s that share very few species with the target vegetation unit Frequency Positive Fidelity Index FPFI is a combination of the Frequency Index and Positive Fidelity Index FPFI ror PFDI Eq 17 86 The index is defined in an interval 0 100 For the most part it retains the advantages and lacks the disadvantages of both This index is preferred for assignment of a relev to a vegetation unit due to its robustness with respect to the heterogeneity of the data set Advanced information about these indices is available in the published paper Tichy 2005 2 4 2 The use of similarity ind
35. exceeds the frequency expected under the null hypothesis of independent distribution of these species in the dataset In many cases however it is better to set this minimum value to be half of the species in the group The minimum value can be specified manually if the Manual setting box is checked 5 Choose a relev colour and press the Select relev s as button to recolour relev s where at least the minimum number of species is present 1 e where the group is present 6 Press the Fidelity Calculation button to calculate fidelity for the coloured group of relev s This opens the Interspecific Associations window Interspecific Associations x Group Positive association Number of relev s Total dataset ba i 155 27 Maximum fidelity as eei ats 43 39 Mark species in the table Pee g 41 39 39 51 38 03 Add to RTF file 36 48 36 31 Wi Only selected species 35 07 35 70 35 59 91 53 Negative Mark species in the table 730 10 8 Carex pediformis s lat 1726 47 S Caragana pygmaea uk x 26 29 8 Poa sect Stenopoa 22 57 i 22 40 21 78 21 43 20 62 20 46 19 35 18 87 i 17 28 Acl ric 9 Fig 72 Calculated fidelity for a selected species group which consists of two species The discontinuity of the species list indicates that this group must be rejected 80 7 Select any new species whi
36. function works only in Similarity Index Display mode Selecting it when the table is in Standard Display mode causes JUICE to crash 2 5 Beals Smoothing Beals smoothing Beals 1984 McCune 1994 replaces a target species s presence absence in the community data with a probability pj of species occurrence in that particular site based on the joint occurrences of the target species with the species that actually occur in the relev site bise BSS Y NC Eq 18 i keR where S is the number of species in relev i Nj is the number of joint occurrences of species j and k N is the number of occurences of species k and the sum is taken over all species that actually occur in relev i all k in Rj 88 This transformation is a smoothing operation designed for community data McCune 1994 As with any numerical smoothing it tends to reduce the noise in the data by enhancing the strongest patterns in the data Beals smoothing can be a time consuming operation 2 5 1 The Calculation of Beals Smoothing in the Program The Beals Smoothing function is called from the Analysis menu Values can be calculated for either the selected relev or for all relev s of the selected colour The co occurrence probabilities can be based on all relev s in the table or all relev s of the currently selected colour For Beals smoothing of a single relev the window shown in Fig 78 appears after selection of the relev
37. in a block or to deselect a currently selected species All information can be saved into the current rich text format export file See Section 1 9 2 Pressing the Export button opens the following window Export Of Diagnostic Constant And Dominant Species Diagnostic species Threshold value 60 Threshold value Bold text style gt Constant species Threshold value 60 Threshold value Bold text style gt Dominant species Threshold value 0 Threshold value Bold text style gt Compact form Cluster arex canescens 6 100 0 6 70 4 aerulea 6 82 3 Alnus glutinosa oe 100 Betula reis 1 Carex cespitos Ja 1 33 arundinacea caerulea 6 100 33 Fre Export Clusters 1 8 Fig 58 Window for Export Of Diagnostic Constant And Dominant Species The list can be formatted as unsorted sorted alphabetically or within layers with or without frequencies or fidelities with or without layers and in compact or extended form The Lower threshold values are fixed they have been defined in the previous window The Higher threshold values Bold text style correspond to the values set in the Synoptic tables tab They may be reset to any value equal to or higher than the lower threshold Changing the Bold text style value for diagnostic or constant species in this window has the same effect as changing the Higher fidelity or frequency threshold value in th
38. maximum likelihood estimation of Beta and HOF response models URL http cc oulu fi jarioksa softhelp hof3 pdf Pielou E C 1975 Ecological diversity John Wiley and Sons Inc New York NY Pl va K amp Pr a E 1969 Typologick podklady p stov n lest St tn zem d lsk nakladatelstv Praha Podani J 2001 SYN TAX 2000 Computer programs for data analysis in ecology and systematics User s manual Scientia Publ Budapest R Development Core Team 2005 R A language and environment for statistical computing R Foundation for Statistical Computing Vienna URL http www R project org Schr der H K Andersen H E amp Kiehl K 2005 Rejecting the mean Estimating the response of fen plant species to environmental factors by non linear quantile regression J Veg Sci 16 373 382 Smith B amp Wilson J B 1996 A consumer s guide to evenness indices Oikos 76 70 82 Sokal R R amp Rohlf F J 1995 Biometry 3 edition W H Freeman and Company New York ter Braak C J F amp Smilauer P 2002 CANOCO reference manual and CanoDraw for Windows User s guide software for canonical community ordination version 4 5 Microcomputer Power Ithaca New York ter Braak C J F amp Looman C W N 1986 Weighted averaging logistic regression and the Gaussian response model Vegetatio 65 3 11 Tich L 2002 JUICE software for vegetation classification J Veg Sci 13 451 453 Tichy L 2005
39. o Sorting Sort species Black Red Blue Sea green Green Yellow Violet Grey All oloooooooo By relev s Red Blue Sea green Green Yellow Violet Grey All 26 Sorting oloooooooo Sort relev s Red Blue Sea green Green Yellow Violet Grey All oloooooooo By species Black Red Blue Sea green Green Yellow Violet Grey All E Sort only by number of species records in releve Ignore ranks of species and cover values Cancel Fig 22 Windows for sorting species and relev s Cancel Note The relev Sorting window has a check box for altering the sorting method When the box is checked only frequency matters and the order of relev s with the same number of species records will not be changed 1 6 3 2 Sort Species Alphabetically can sort all species into alphabetical order or it can be restricted to only sort the species of the selected colour Other Species Sorting Functions Sort Species By Species Data sorts according to the information stored in the Species Data Column See Section 1 7 for information on how to write data to this column This column can contain many types of information such as layer Ellenberg indicator value frequency or any other biological information about the species The Species Sorting Parameters window has several options The sort can be restricted to species of the selected colour or expanded t
40. pendula 1 100 Alnus glutinosa 1 70 4 Juncus effusus 6 100 Anemone nemorosa 6 33 Frangula alnus 4 69 9 Alnus glutinosa 4 100 Alnus glutinosa 1 33 Betula pendula 1 63 7 Dryopteris carthusiana 6 67 Sorbus aucuparia subsp aucuparia 7 33 Alnus glutinosa 4 67 Melica nutans 6 67 Frangula alnus 4 67 Carex cespitosa 6 Include diaanostic species ae E Show references Minimum req 1 100 o LJ m SAR DeC Proc cOP A 2p P a Mac eu ae a ee RNC Cr RE i Mark in the table black Mark in the table black v Mark in the table black Aver species No a s Aver positive fidelity Sharpness Column a Fig 57 Window for detailed analysis of synoptic tables From the Synoptic Table menu select Analysis Of Synopt Columns Initially the three lists will appear blank Select the column to analyse using the slide bar at the bottom and then press the Refresh button Three species lists will appear Diagnostic species are those with fidelity higher than the Lower fidelity threshold defined in the Synoptic tables tab of the Options menu see Section 1 11 1 Constant species are those with relative frequency higher than the Lower frequency threshold also defined in the Synoptic tables tab However only constant species that are not diagnostic will be included in the list unless the Include diagnostic species
41. saved in the file result_table csv in the directory of the R package usually the folder c Program Files R R 2 2 1 bin and can be opened e g in a spreadsheet program for other analyses The list of species that can be analysed together is limited to 10 The calculation of SRC for all species in the data set functions similarly With so many species there is no graph and the function produces only a table result_table csv in spreadsheet format 96 Literature Austin M P 2002 Spatial prediction of species distribution an interface between ecological theory and statistical modelling Ecological Modelling 157 101 118 Barkman J J 1989 A critical evaluation of minimum area concepts Vegetatio 85 89 104 Beals E W 1984 Bray Curtis ordination an effective strategy for analysis of multivariate ecological data Adv Ecol Res 14 1 55 Botta Dukat Z amp Borhidi A 1999 New objective method for calculating fidelity Example The Illyrian beechwoods Ann Bot 57 73 90 Botta Dukat Z Chytry M Hajkova P amp Havlova M 2005 Vegetation of lowland wet meadows along a climatic continentality gradient in Central Europe Preslia 77 89 111 Bruelheide H 1995 Die Grinlandgesellschaften des Harzes und ihre Standortsbedingungen Mit einem Beitrag zum Gliederungsprinzip auf der Basis von statistisch ermittelten Artengruppen Diss Bot 244 1 338 Bruelheide H 1997 Using formal logic to classify vegeta
42. the No of iterations of this random selection and calculation The resulting values are the averages over all iterations These are saved to the Clipboard Note The Data Transformation field is active only for calculation of Euclidean distance All other measures are based on presence absence data 2 7 Euclidean Distance Euclidean distance is the most intuitive and common method of distance measurement between two sets two relev s a relev and a constancy column etc Mathematically we compute the differences along each axis sum the squares of the differences and take the square root of that sum Eq 19 where x is the cover value of species i in the first relev y is the cover of species i in the second relev and n is the number of species in the table 2 7 1 How Euclidean Distance Is Calculated Euclidean Distance is included into two functions Total Inertia Euclidean Distance Beta Diversity from the Analysis menu and Euclidean Distance from Selected Species from the Head menu with the Store Values to Short Headers function The first of these calculates both beta diversity and Euclidean distance for a selected group of species and the operating window is the same as in Section 2 6 1 The process averages all Euclidean distances 90 between two pairs of relev s The list of relev s may be selected by colour or by separators The user may standardise each group to the s
43. the cursor is in the species names Left Button Click Left Button Double Click Left Button Click and Drag Shift Left Button Click Right Button Click Ctrl Right Button Click Shift Right Button Click Move currently selected relev group column Highlight and select species Open dialog window for editing species name layer and data Move currently selected species Make remove separator line below currently selected species Repaint currently selected species with current species colour Repaint currently selected species with current secondary species colour Repaint block of species with current species colour Click on the first species to be selected Hold Shift and click on the last species The entire species interval will be repainted 13 14 In header data Relev s Head Sorting Separators Synoptic Table Indicator Values Analysis Table Simulation Help Short head Relev number Month Altitude m Biblioreference Day Aspect degrees No table in publ Author code Slope degrees No relev in tab Syntaxon code Mosses ident Year Relev area m2 Locality 592 Row 1 Achillea aspleniifolia Cokin Fig 8 Header data These functions are defined for Header Data Display When the cursor is in the header data epaint block of relev s with current relev colour Clic on the top relev to be selected Hold Shift and click on the bottom relev The entire relev interval will b
44. the mouse in combination with the keyboard 1 5 2 1 Functions sorted by displayed objects In tables JUICE c juice_fin 1 louky_ass_str_del01 wct DER Fie Edit Species Relev s Head Sorting Separators Synoptic Table Indicator Values Analysis Table Simulation Help IP LT PS Te PS E CC EE ep Gm m Statistics _ U value hyp jad ij Total time 2 days 5 h 39 min 21 sec Running number T 111 11 1 766 8819555 5552 5749978 77 733311070189777788 208333333377777 Releves 1210 688 5964999488053864995229919222947374224499441674111121112222 Species 606 765434639210087139933236134965567207283206576237864123415690123 Achillea aspleniifol S Achillea millefo Achillea ptarmica Acinos arvensis S Aconitum napelus Adoxa moschatellina Aegopodium podagrari S Agrimonia eupato Agropyron caninum S Agrostis canina Agrostis capillaris S Agrostis stoloni Ajuga genevensis Ajuga reptans S Alchemilla hybri S Alchemilla vulga Alisma gramineum Alisma lanceolatum Alisma plantago aqua Allium angulosum Allium carinatum Allium oleraceum Allium scorodoprasum Allium vineale Frequency 440 Rel o Turboveg No 11 Agrostis capillaris 6 Fig 6 Table 12 The three parts of the table see Section 1 5 1 above are sensitive to different operations When the cursor is in the short headers Left Button Double Click Display list of species in selected relev and save selected relev in text f
45. which a large proportion of diagnostic species of vegetation unit j are also diagnostic species of vegetation unit k It is an asymmetric similarity measure because in cases of low numbers of diagnostic species in vegetation unit j and high numbers in k where most of the diagnostic species of j are shared with k the similarity of j to k is high while the similarity of k to j is moderate The Similarity Index is calculated separately for all pairs of classes and all pairs of alliances suballiances For every pair j and k two indices are calculated one for the similarity of j to k and the other for the similarity of K to j Second the Uniqueness Index for each vegetation unit is calculated using this formula 1 xe Eq 9 J 3 Eq 9 This Uniqueness Index is low for those vegetation units whose diagnostic species are mostly shared with other vegetation units 1 11 13 2 Uniqueness in JUICE The Uniqueness function is available for synoptic tables with fidelities After selecting Uniqueness from the Synoptic Tables menu the following window will appear 64 Uniqueness tj Group No 1 Uniqueness for the fidelity cut level 10 0 591789439018335 Assymetric simil index Assym SI Species Average Group No 1 15 3 Group No 2 17 5 Group No 3 2 25 9 Group No f 33 3 Group No 2 33 1 Group No 33 3 Group No A 34 3 Group No 32 5 Unit 41 TT Fidelity cut level Calculate one column Calc
46. within each group of relev s Groups are defined by separators See Section 1 5 4 From the Head menu select Short Header Averages The window shown below will appear By default the average value of the short header across each group is displayed There are option buttons for displaying the average the minimum or the maximum The list of values is also copied onto the clipboard It can be pasted into a text editor with the command Ctrl Insert or Ctrl V 40 Average L Minimum Maximum List is copied to the clipboard Fig 36 Averages minima and maxima from short headers calculated for each constancy column of the table 1 8 5 Short Header Sorting It is often desirable to sort relev s according to some criterion First the values under consideration should be written to the short headers as described in Section 1 8 2 Then from the Sorting menu select Sort Short Headers For more information on sorting see Section 1 6 3 especially Section 1 6 3 3 1 9 Exporting Data JUICE can produce several types of data phytosociological tables synoptic tables ecological information about relev s or species etc To make this information available for use by other programs JUICE supports many types of exports 1 9 1 Saving Files in JUICE Once the source table data are imported into JUICE they can be saved in JUICE s format which may include extra parameters From the File menu select
47. 0 4 M T T T T T HOF optimum 4 4056 4 5 6 7 8 HOF min 3 70 HOF max 5 2239 GRADIENT HOF prob optimum 0 7609 HOF interval 1 5239 HOF model Vv Fig 83 Window for calculation of species response curve for one species The window contains a graph and a table giving optimum minimum maximum interval and the probability of the optimum value for each model graphed 95 Note All table values are saved in the file result_table csv in the directory of the R package usually the folder c Program Files R R 2 2 1 bin Graph r3 Displayed values e Optimum C Minimum C Maximum HOF C Interval C Optimum prob fi fi 1 1 fi Values Polytrichum strictum 37 2 08 4 Ree H Pleurozium schreberi 37 D g Sphagnum girgensohnii 4 1704 2 Diplazium sibiricum 5 4444 8 06 7 H o 5 04 ee L o n x NCC DEP TN 9 024 CE x H a P d BS T M n d 2 4 AC Eo 0 0 71 cilm r T T T T T 4 5 5 7 8 GRADIENT Diplazium sibiricum Polytrichum strictum Pleurozium schreberi Sphagnum girgensohnii 7777 Close Fig 84 Window for calculation of species response curves for multiple species The calculation of SRC for multiple species allows easy comparison of species optima maxima minima and their intervals The list of species on the right side of the window see Fig 84 is sorted in ascending order according to the selected value Note All values are
48. 2 Visual Basic tc Corel PHOTO P Total time 3 days 13 h 0 min 25 sec cs Fig 49 Synoptic table with categorical constancy Statistics gt JUICE c documents and settings ubomir tichy dokumenty 0_lubos juice kurz 10_3_2005 dyje_valley1 wct File Edit Species Sorting Synoptic Table Table Simulation Help Absolute frequency synoptic table Number of releves releves 202 Species 631 Poa trivialis Deschampsia cespitos Equisetum arvense Aegopodium podagrari Urtica dioica Geum urbanum Stellaria nemorum Galium aparine Pulmonaria obscurato Sambucus nigra Galeobdolon montanum Phalaris arundinacea Oxalis acetosella Carex sylvatica Glechoma hederaceath Carex remota Dryopteris filix mas Carpinus betulus Hepatica nobilis Dactylis polygama Quercus petraea agg Fig 50 3 FP 00HPU QU0 U10oY6 0 0 0 ere DPUNMMMM N WN NEBNNND i l l EN I doce Ea 4 Total time 3 days 13 h 1 min 43 sec WRU OU OPW Ne Synoptic table with absolute constancy S amp 27 12 09 1 11 4 Synoptic Table with Absolute Frequency Absolute Constancy 56 In this synoptic table the total number of relev s in the relev group is displayed at the top of each column The entries in the table are the number of species occurrences in the relev groups The highlighting scheme is the same as for percentage synopti
49. 3 4 or 5 degrees of freedom 4 Huisman Olff Fresco models HOF a hierarchical set of five models with increasing complexity The bell shaped curve is considered mainly for reference The other three options allow flexible expression of different response curve shapes each having different constraints advantages and disadvantages GLM models offer curves which are completely described by an equation with a given number of parameters however the shape the curve is quite often inappropriate or unrealistic GAM models are more flexible in terms of curve shape but their equation is non parametric and not easily expressible Perhaps the best option from the ecological point of view is HOF models which have properties one would expect from species response curves For example they express only unimodal response shape which is consistent with the assumption that a species has a single optimum condition along an environmental gradient GLM with a low degree polynomial and GAM with few degrees of freedom cannot really produce bimodal response shape but they can easily produce a semi bimodal shape with inexplicit interpretation Available information about species is reduced only to presence absence even though all methods can handle percentage data and transformed data The curves resulting from presence absence data are more aesthetic yielding shapes with a more straightforward interpretation Also removal of information about dominance from s
50. 7925227048566981174536586 GROUP Rhinanthus major 1 72 1 Total 18 006600100800000000000000000616000066000 1 Rel o Turboveg No 2 Rosa canina s lat Fig 30 Species group tables 34 1 7 9 Species Data Averages Groups of species defined by separators can have different species data values If these data are numerical it is possible to calculate averages which can reveal differences between species groups To view averages of species data open the Species menu select Species Data and Species Data Averages 1 7 10 Species Data Exports Species data displayed in the Species Data Column can be saved into a simple text file which can then be imported into other tables as described in Section 1 7 7 From the Species menu select Species Data and Export Species Data For more information see Section 1 9 8 1 8 Short Headers Short headers can contain up to six characters displayed vertically above the table data Not limited to identification numbers this field can contain any brief information about a relev that could be useful for relev identification classification or sorting Various functions for specifying what is displayed in the short headers are found in the Head menu Relev s can be sorted according their short headers as explained in Section 1 6 3 3 1 8 1 Identification Numbers The Head menu has functions for displaying any of four dif
51. 995 According to the central limit theorem e g Zar 1999 76 77 if the number of species is high which is the case in most data sets the crispness of classification has approximately a normal distribution with expected value c 1 and standard deviation 2 c 1 S where S is the number of species The effect of the number of clusters is removed by subtracting this expected value and dividing the difference between the observed and expected value by the standard deviation In this way the crispness values are standardized and can be compared among partitions with different numbers of clusters Local maxima of crispness may indicate optimal numbers of clusters 1 11 15 2 Optimal Use of This Function in JUICE This function is used to its full power in conjunction with the function Sort Relev s By Clusters see Section 1 6 3 3 1 Call the function Sort Relev s By Clusters Open the resulting file produced by PC ORD or MULVA or classify the data set using the Cluster Analysis PC ORD function from the Analysis menu The sorting window will appear automatically after the analysis This window is always displayed on top but the table remains active 2 From the Synoptic Table menu select a table with frequency or fidelity values Repeated selection of different numbers of clusters will automatically change the displayed synoptic table 3 Choose a species colour and use it to select a list of species a
52. AND lt Mercurialis perennis OR lt Urtica dioica gt Before running the query its syntax can be checked using the Show definition button This displays the hierarchy more clearly lt Ulmus glabra gt OR lt Carpinus betulusUP25 gt AND lt Mercurialis perennis gt OR lt Urtica dioica gt Warning All pairs of logical variables connected by one operator must be put in parentheses 2 3 7 Expert System The Expert System function of the Analysis menu can automatically assign a relev to a vegetation type if there is already a classification based on species groups The classification algorithm must be included in a TXT file This file should preferably be created as a product of the classification in a large data set and must include all required information on aggregated 83 species 1 part species groups 2 part and their combinations into vegetation types 3 part If the user also wants an a posteriori classification of relev s by similarity indices the file must contain a synoptic table with constancy columns for all vegetation types 4 part The file has a text structure and can be created manually In the first part of this file merged species are defined species name and layer The second section defines species groups The third section contains logical formulas with community definitions 84 The first line of each definition contains its hierarchical l
53. E Fig 39 Example of the distribution map created by D MAP 1 9 8 Species Data Export Species data see Section 1 7 imported or created manually see Section 1 5 7 can be saved as a text file for future use These data can then be imported into similar tables The export file has the following structure species name layer and species data value Fallopia convolvulus 6 arc Lapsana communis 6 arc Silene latifolia subsp alba 6 are Viburnum opulus 4 nat Sorbus aucuparia subsp aucuparia 4 nat Ranunculus sceleratus 6 nat Populus x canadensis i neo Epilobium ciliatum 6 neo Pimussa i neo 1 9 9 Short Header Export Header data can be exported into a simple text file for later retrieval See Section 1 8 2 9 The file contains the relev number short header value group number and a 1 or 0 representing presence or absence of a separator to the right of the relev 400001 40 1 0 400002 Woe Al 0 400003 mes d pod 400004 Sep 2 2 0 400005 4l s 2 m 0 400006 wise Z2 p d 400050 922 3 Qj 400051 39g 3 g 0 400052 asp S e di 400053 52g d 2 0 400054 ise A p 400055 562 E 45 1 10 The Fidelity Concept Chytry M Tichy L amp Holt J Statistical fidelity measurements are useful for measuring species concentration in vegetation units and comparing diagnostic values among species in a particular vegetation unit or among vegetation units for a particular species A completely statistical approach to determinat
54. ICE c uice finMlouky ass str del01 wct File Edit Species Relev s Head Sorting Separators Synoptic Table Indicator Values Analysis Table Simulation Help STA vum 159 Total time 2 days 8 h 11 min 43 sec Tae pa je rn we IEE Jes Statistics U value hyp jad Running number Releves 1210 Species 606 Holcus mollis Homogyne alpina Humulus lupulus Hydrocotyle vulgaris S Hylotelephium ma Hypericum hirsutum Hypericum maculatum Hypericum perforatum Hypericum tetrapteru Hypochaeris maculata Hypochaeris radicata Impatiens noli tange Impatiens parviflora Inula britannica Inula hirta Inula salicina Iris pseudacorus Iris sibirica Juncus acutiflorus Juncus articulatus Juncus atratus Juncus compressus Juncus conglomeratus Juncus conglomeratus 766 88 555 4 107018977 337195552 688 5948803414737422491112649995 765434608719360728320677340392103 Releve No 1163 Turboveg No 587205 Edit header No of species 20 20 records Shannon Wiener Index 2 04 Equitability 0 68 Layer Cover C Aph C Seq S Avenula pubescens Anthriscus sylvestris S Festuca rubra agg S Dactylis glomerata agg Arrhenatherum elatius Lupinus polyphyllus S Achillea millefolium agg Cirsium arvense Trisetum flavescens Alopecurus pratensis S Galim album s 1lat S Rumex acetosa Poa trivialis Geranium pratense S Anthoxanthum odoratum S Galium verum agg Hypericum perforatum S Trifolium repe
55. JUICE program for management analysis and classification of ecological data Tichy Lubomir amp Holt Jason VEGETATION SCIENCE GROUP MASARYK UNIVERSITY BRNO Czech Republic 2006 Program manual Dept of Botany Masaryk Univ Brno Kotlarska 2 611 37 Brno tichy sci muni cz Box 37 Hinsdale MT 59241 USA jholt seznam cz Table of Contents 1 GETTING STARTED wr cscscccvcseccacvcscnsadustcvcacvatenssdvetcusscvescuccavetcueacvetcuosdvebcuscaveacueads 1 1 1 IntroQUCO0n eire terere sescsciens DS geeks oe eve etus yere einge sousdcessesusdeckncssnestenssssuessuiessdessennsdoseadeusnosonsstes 1 1 2 Copyright Information 1 1 3 Ts tall ation PAE EEES EE EEEE TE EE T soescunnss Sessensesoscesesusssowasees 1 1 3 1 Content of the Installation Package eese eene enne nnn 1 1 3 2 Computer Configuration i2 tete ee do epo OO eie eiit ES 2 1 3 3 Program Settings and INI File eere en nnns 2 1 4 lirirdbiio qe D 3 1 4 1 XMELE Fottnat nece D RB e e PIS D ERES 3 1 4 2 Cornell Condensed CC Format sess eene nennen nren nennen enn 3 1 4 3 suu d ME 5 1 4 4 Spreadsheet Format decode recede e GRE ee e e EE te e Toe Y e Heres 7 1 4 5 Text Format ze eee e eee ede es 7 1 4 6 Header Data inii reri e e E UTER RET UI HE e DEED e een eas 8 1 5 The Basics of Working with Tables ceres scene ente eene en nennen neta seta
56. N N N Eq 7 This formula can be easily extended to a comparison of vegetation units within the classified data sets by replacing the species frequency outside the target vegetation unit n n ON N in the denominator by the sum of relative frequencies in all the other vegetation units of the given typology The Dufr ne Legendre Indicator Value is independent of the relative size of the target vegetation unit It implicitly standardizes the size of all vegetation units in the data set including the target vegetation unit All vegetation units are weighted equally and changing N does not influence the resulting numerical value The categorical form of the Dufr ne Legendre Indicator Value probably deviates from most researchers intuitive expectations of the properties of a suitable fidelity measure The next figure shows that the Dufr ne Legendre Indicator Value gives a rather high weight to common species It could be said however that the Dufr ne Legendre Indicator Value is a fidelity measure which emphasizes the group s fidelity to the species 1 e the value is a good indicator of how frequent the species is within the vegetation unit but it is less affected by occurrences of the species outside the vegetation unit 0 0 2 0 4 0 6 0 8 n no N Np Fig 41 Dependence of the Dufr ne Legendre Indicator Value on the relative frequency of species occurrences within vertical axis and outside horizontal axis the targe
57. Perc Co occurrence 206 species Number of relev s gB TSE ocomium s endens Total dataset 311 y P 91 30 Stellaria bungeana COM oo 00 Phytidiadelphus triquetrus ETT 82 60 Dryopteris expansa 82 60 S Abies sibirica 82 60 S Spiraea chamaedrifolia hieck gl 78 30 Oxalis acetosella J 178 30 Plagiochila porelloides 78 30 Cruciata krylovii a 73 90 Calamagrostis obtusata E Only selected species 173 90 Gymnocarpium dryopteris d 69 60 Diplazium sibiricum 69 60 Pleurozium schreberi Fig 68 Co occurring species window The selected species is shown in the upper left corner with the number of relev s in which it occurs below The list of the most frequently co occurring species is sorted by decreasing frequency in the relev s where the selected species occurs The value shown in the first column is the percentage of relev s of the selected species also containing the listed species The next columns are species name layer species frequency in the data set and frequency of joint occurrence of current and selected species Note Species listed in the window list may be selected and marked with a colour which will appear back in the table The list can be saved into the previously defined RTF file as described in Section 1 9 1 2 3 2 Interspecific Association Interspecific association between the selected species and each other species in the table is the basic step of the COCKTAIL method A fideli
58. Species Statistics again Sort Species by Average Vegetation Richness is similar to the previous function Average vegetation richness AVR is calculated by averaging the number of species in all relev s in which the species occurs AVR is not displayed after sorting To view AVR and other species information select Species Statistics from the Species menu To return to Standard Display mode select Species Statistics again Dependence Sorting uses interspecific associations as the main sorting criterion Interspecific associations between all possible pairs of species are calculated according to the selected fidelity index For each species the average value of the selected fidelity measure of the most similar species is calculated and the data set is subsequently sorted by these average values The number of similar species considered can be 1 3 5 10 or the calculation can include all the species that occur in any relev with the species in question Dependence Sorting Average is Calculated from Sort species Black the most dep species MDS Red first three MDS Blue first five MDS Sea green first ten MDS Green All common occurences Yellow Minimum frequency di sorted Violet i i gt 4 Grey oO o Oo o o o o o Expected processing time All 00 00 10 Cancel Continue Fig 24 Sorting species by maximum fidelity values Warning 1 Check the estimated
59. Text Format For users without TURBOVEG this is the easiest import format The table consists of four files with the same name and different suffixes TXT TAB EXP and STR Similar files can also be exported by JUICE From the File menu select Export and Table and choose the TEXT FILE format The TXT file consists of three columns species name layer number and table data separated by at least 5 spaces Anemone nemorosa Athyrium filix femina Atrichum undulatum Avenella flexuosa Betula pendula Betula pendula Calamagrostis arundinacea Calamagrostis villosa Carex leporina Deschampsia cespitosa Dryopteris dilatata Epilobium angustifolium Equisetum sylvaticum Fagus sylvatica Fagus sylvatica 22982 e Sra S Scocdbs 544533 4 H oO00050000540t00 0 The TAB file is optional Each line contains a unique relev identification number in a range from 1 to 999 999 434111 434112 434113 SUMI SHELL 311724 The EXP and STR files contain header data See Section 1 4 6 1 4 6 Header Data There are three ways to import header data into JUICE 1 Header data are included automatically in the XML file exported from TURBOVEG You can select which fields to import during the import of this file into JUICE as described in Section 1 4 1 2 When TURBOVEG exports a CC file for JUICE it automatically exports header data in an STR and an EXP file The user must select header data fields to export durin
60. able unification of the species nomenclature used for defining species groups It is possible to classify either a single relev causing its species groups and Cocktail classification to be displayed or all the relev s of the currently selected colour causing the codes of their Cocktail groups to be written to the Short Headers field 2 4 Similarity Indices The main purpose of classifying vegetation is to organize ecosystems into groups that are similar in their floristic composition and structure Calculating similarity indices between new relev s vegetation plots and constancy columns of vegetation units is a suitable method for assigning the relev s to these units It is one of several possible approaches that can be formalised and included into the computer expert systems used by environmental and nature reserve managers This method was firstly described by Hill 1989 who matched relev s to constancy columns in the tables of the British National Vegetation Classification using an index of compositional satisfaction which was derived from the Czekanowski coefficient Czekanowski 1913 85 2 4 1 Description JUICE supports calculation of three similarity indices Frequency Index is able to measure the compositional similarity of relev s while the other two indices Positive Fidelity Index and combined Frequency Positive Fidelity Index also include information on diagnostic species which is a very important syntaxonomic characte
61. al ues dle K i 412712 472632 iie der 15 407248 400153 0024 21BE02 458623 458502 0021 21BE07 00 540 404645 400139 0021 21BE07 6 00 490 407282 400153 0024 21BE06 00 404648 400139 21BE07 00 470 423759 400612 21BE06 407255 400153 21BE02 407143 400153 22AB01 404643 400139 21BE07 407150 400153 22AB01 407264 400153 21BE02 458539 458502 21BB06 403818 400117 22AB02 N kfi 2 km Bezdekov p Blatn Aue Mahou 1k Sedlice sti Zelezn hor Cejkovice Nakri 2km Zwischen Bl Vlhlavy lk Albrechtice Sudlicher R Louky JZ od ooooooooooooo ooooooooooooo Frequency Rel No 71 Row 1 S Plagiomnium affine agg 47132 Turboveg No 407282 Column Fig 18 Header data display 22 All screen parameters are defined similarly as in the Standard Display See Section 1 5 8 Mouse functions are described in Section 1 5 2 If it is necessary to change this data the changes should be made in the source data set However it is possible to edit this information within JUICE See Section 1 5 7 Note It is not possible to add a new header field to the header data To add new information define the new field in TURBOVEG and re export the header data See Section 1 4 6 1 5 10 1 Selecting Relev s by Header Data Section 1 5 5 explains how to organise colour coded relev s into blocks JUICE includes functions for colour coding relev s according to their header data 1 Select a relev colour on the Icon
62. alue is influenced by the turnover of species among habitats Koleff et al 2003 JUICE includes such measures as Whittaker S rensen dissimilarity Jaccard dissimilarity Harrison beta and Williams beta calculated from presence absence data 2 6 1 How to Measure Beta Diversity in JUICE The program can make calculations according to any of the most widely used beta diversity measures Users may call this function from the Analysis menu by selecting Total Inertia Euclidean Distance Beta Diversity This opens the following window Fig 79 89 Total Inertia Euclidean Distance Beta Diversity Colour Grou 74 212 No ofrelev s 10 M No ofiterations 10 Data variability Beta diversity only pres absence Average Total Inertia 2 9165 Whittaker overall 2 4361 Average Euclidean Dist 52 3707 Jaccard dissimilarity 0 6980 Sorensen dissimilarity 0 5546 Harrison beta 2 0 3331 Williams beta 3 0 2370 Close Fig 79 Calculation of beta diversity and Euclidean distance for a selected part of the data set The group of relev s from which beta diversity is calculated may be selected either by colour or by separators The measure is comparable between columns in cases when the number of relev s in each relev group is the same For this reason the user has the option to limit the number of relev s selected for analysis The relev s analysed are selected randomly The user can specify
63. alues must be specified manually All codes which have been found in the table will be displayed in the list on the right side of the window Double click on a code to enter a percentage value for the cover represented by the code The program can continue only if all codes are associated with a whole percentage value An example is given below An example is also available on the JUICE web page Table from relev s of the file paseky wct Number of relev s 5 eee iy epee a ek Tee ete ou ree ey Calamagrostis villosa 6 5 4 4 5 3 Avenella flexuosa 6 2 2 3 1 2 Wekee akiaakionn omoia e INOS Pal gal eal ail er Rubus aC OAD ODO 2 19 Maianthemum bifolium 6 2 3 3 Veratrum album ssp lobelianum 6 2 2 3 OxalklsmaceroceMar iG aii 9 9 Senecio fuchsirrorlgvwb fied Beas ete ms valliviadts CUm A172 2 e M ze Phegopteris connectilis 60 2 2 f 1 Anemone nemorosa 6 2 2 Becuo pendul ario reran Piecsa Ppungensr Asar ei A22 Becule oemculap op 5p opp Galium saxatile 6 2 Trientalis europaea or ACHTMANN REE eiO rr caress Deschampsia cespitosarolt tort re dias Prater Gee te Wu ome pe The second column represents layer it is optional All cover data can be converted into percentage numbers or some semi quantitative scale This format can use full species names or abbreviations The spreadsheet format must not contain header data Header data can be imported separately as described in Section 1 4 6 1 4 5
64. alyse the data outside of JUICE and import the results by doing the following 1 Export the data set to a CC file Run MULVA Import the CC file into MULVA Apply data transformations Select the parameters of the cluster analysis and analyse the table From JUICE s Sorting menu select Sort relev s by clusters and MULVA 7 Specify the location of PRINDA the MULVA output file It is then possible to proceed as described in Section 2 2 1 3 eon d DS 2 2 3 Cluster Analysis via SYN TAX The SYN TAX 2000 package Podani 1999 also contains programs for data analysis The HierClus module calculates hierarchical clustering This software is commercially distributed and available e g at http www exetersoftware com cat syntax syntax html 2 2 3 1 Manual Table Analysis and Import of Results At present it is not possible to run SYN TAX from within JUICE To manually use SYN TAX cluster analysis with JUICE do the following 1 From JUICE s File menu export the table by selecting Export and SYN TAX input files 2 Run SYN TAX 2000 s HIERCLUS EXE module 3 Import the data file into SYN TAX 4 Select the type of cluster analysis data transformation and resemblance function 5 From SYNTAX s File menu select Save output files to open the Save files window Select switches Tree and Sequence of classified objects 6 Analyse the table 7 From JUICE s
65. alysis 5 ereeyee ea tore veio rete eno nier eret aa sunsdsessedosescecncsnnseesess ss ee Ewa Pal E eee Ta or o UN EPVS ee EYa S Ue roue 69 2 2 1 Cluster Analysis via PC ORD nee eee eerte eR e REED ee ve e b e ue Ce Run 69 2 2 1 1 Initiation of the Connection o Rx AO AR eoa REP vo er E Hee rS 69 2 2 1 2 Automation of Cluster Analysis Within JUICE esee 70 2 2 1 3 Results and Cluster Trees uae Abie seine Re OE rer ere ipta eet 70 2 2 1 4 Manual Table Analysis and Import of Results essere 72 2 2 2 Cluster Analysis via MUL V A hne se ee Ue e SERT NR EO OUR TH RE neh ket 72 2 2 2 1 Initiating Connection etes ates is e NU tee E B REDEUNT 73 2 222 MULVA Cluster Analysis from within JUICE sees 73 2 2 2 3 Results and Cluster Tr6e e SERRA RR tes RR A E A IRR eno a Tia 73 It is also possible to analyse the data outside of JUICE and import the results by doing the following 74 2 2 3 Cluster Analysis via SYN TAX 0 ccccecsseescesscessceecesecsecasecaeecaeeeseeseeseeeeeeeseceseceaecaecaecseeceeeneeaes 74 2 2 3 1 Manual Table Analysis and Import of Results sse 74 2 3 COCK TATE Method eere eoru eI o eee e ee oe None ge oe ee deoa oe kg poe Eg e oe e nee on NS 74 2 3 1 Co Occurring Species seem e ave tee EIN uc epis RHET TRIER cones 76 2 3 2 Interspecific Association aceite dee ie ete te ee iiei i RR eeii 76 2 3 3 Initial Selection of Species into Groups
66. am can calculate crispness of classification with a limited list of species It is recommended e g to exclude rare species such as those occurring in fewer than 10 relev s from this analysis Note 2 It is also possible to search for maximum crispness using hierarchical separators See Section 1 5 4 Varying separator level will change the number of columns in the synoptic table This can be useful e g with results from TWINSPAN classification 1 11 16 Comparison of Two Synoptic Tables If the synoptic table is displayed in the form of percentage frequency or fidelity it is possible to compare two synoptic tables and find their total similarity This function was used in the paper Knollova et al 2005 for comparison of several different classifications The program calculates a crossing table of Euclidean distances between all pairs of synoptic columns in two synoptic tables Note Calculation of fidelities can be modified by data standardization Fisher s exact test and selection of presence absence versus quantitative data as described in Section 1 10 3 67 1 11 16 1 Theoretical Background The distance d TAB1 TAB2 between the i th column of one table TABI and all columns of a second table TAB2 is calculated according to the formula min ED Mr ap 1 Z ED min ED where i and j denote columns of classified resampled data sets TAB1 and TAB2 respectively ED is the Euclidean distance between ph
67. ame number of relev s by random selection or by bootstrap algorithm It is also possible to repeat the process in more iterations and average the results Species covers can be transformed into square roots or binary presence absence data The second function calculates the Euclidean distance between the selected relev and each relev in the data set and sorts the relev s accordingly thus finding the relev s most similar to the one selected Euclidean Distance from Selected Relev Selected relev Turboveg No 217 Data transformation Click on some relev at the table j N Save ED into short headers EucDist RellD GroupNo ShHead e A h peb peb peb pob pab pab pab pb pb pab jab pab pab AAAA NDA AA AT Aa a wa FON B FAV ROHS 0 ww d into the clipboard Fig 80 Calculation of Euclidean distance between the selected relev and all other relev s in the data set Values can be displayed in the window list or written to the Short Headers Species covers may be transformed into square roots or binary presence absence data All values in the list are copied to the Clipboard Note The last column of the table in the list box is the current value of the Short Header field This field can contain additional environmental data e g pH conductivity average temperature or slope which may be analysed together with the similarity of species composition 2 8 Species Response Curves David Zeleny amp Lubom r Tichy
68. and is loaded each time JUICE is started 1 5 10 Header Data Full header data are displayed separately JUICE must be switched from Standard Display to Header Data Display by selecting Extended Head from the Head menu or by clicking on the Head Icon on the Icon Bar In Header Data Display mode each relev has its own row The header data fields are in the same order as in the EXP and STR files See Section 1 4 6 for more information on these files All names of fields defined in the STR file are written at the top of the table All table data must contain the field Relev number which is a unique identification number connecting headers with table data JUICE c tmp molinion wct BAE File Edit Relev s Head Separators Help AEE A CALARE eiz x E eir K X r z Total time 0 days 0 h 0 min 23 sec head Relev number ac et ub S 3 ENE SERRE MIS aver Year gr layer Author ea lope ATE Mosses ident Syntaxon code Cover tree laye Locality 407232 400153 21BE02 407240 400153 21BE02 407238 400153 21BE02 404629 400139 21BB06 403466 472505 21BE01 407233 400153 21BE02 407252 400153 21BE02 407239 400153 21BE02 407142 400153 Fracht 407242 400153 Pasice 1 k B ehov 0 7 Doman n 1 Sedlice s doln komp Ce ejov bei Svinetice Ostrolovsky Preseka 0 oooooooooo oooooooooo MH HS HH HR HR RH HR 002 30 Y Mahous 0 7 421749 400583 aho v 05 421743 400583 Head av
69. aphical Position Index calculates a virtual grid assigns the same number to relev s in the same square and writes this number to the short headers Such spatial stratification of table data is available if the header data contain information about geographical coordinates The fields LATITUDE and LONGITUDE both 6 or both 7 characters must be included in the header data files with suffix STR and EXP see Section 1 4 6 Example 1 LATITUDE 491357 49 13 57 LONGITUDE 163420 16 34 20 Example 2 LATITUDE 0491357 49 3 57 LONGITUDE 0163420 16 34 20 Selecting this function opens this window Grid Analysis Geographical grid spacing Latitude 7 0 75 Longitude 1 25 Group aggregation support Fig 32 Window for calculation of grid cell relative numbers The window is used to define the size of the grid spacing The default values can be altered according to your project requirements When you press Continue the program will calculate an index for each grid cell and the indices will be written to the short headers The Group aggregation support checkbox is useful when the data are divided by separators When the box is checked each relev group is analysed as a stand alone data set Note The grid indices do not indicate relative position The square containing the first relev is assigned the index 1 The square containing the first relev that is not in square I is assigned the index
70. ated position in the table while the right mouse button deletes the value in the indicated position i e overwrites it with cover value 0 It is best to save a backup file before editing cover data because JUICE does not keep track of the original values so these cover modifications are irreversible After data modification select the Table Simulation menu and Add Remove Species Cover once more and mouse functions will return to normal c Header data can be accessed from the Relev Overview Window See Section 1 5 6 Clicking on the Edit header button opens the following window Header Data Field name Locality No of characters 100 Mahous 1 km SOS von der Gemeinde Fig 14 Editing header data The two arrow buttons enable selection of the field to be edited Click the right arrow button until the desired field name appears Type the new value into the text box Then press the Save button If the Save button is not pressed the value will not change The left arrow button initially does nothing but after the right arrow button has been pressed the left arrow button can be used to go back to a previous field Screen Options Display parameters are defined in the Options window under the Display parameters tab It is possible to modify 1 Displayed length of species names 2 Width of species data field 3 Size and style of text 4 Background contrast Opti ions Fideli
71. bos juice kurz 10_3_2005 dyje_valley2 wct File Edit Species Sorting Synoptic Table Table Simulation Help Emea a cv moons Uu Statistics Phi coett Total time 4 days 6 h 18 min 22 sec Number of releves releves 202 Species 631 Alnus glutinosa 100 Betula pendula 100 Frangula alnus 8T Carex cespitosa Epa Scirpus sylvaticus 3327 Cardamine amara sg n Deschampsia cespitos Aa Carex acutiformis of Carex brizoides Ranunculus repens ed Dryopteris carthusia 100 Crepis paludosa ids Sambucus nigra Aegopodium podagrari Stellaria nemorum Acer campestre Impatiens parviflora Ficaria verna subsp Ulmus glabra Galium aparine Phalaris arundinacea Lamium maculatum zd 3 Carex remota agas Lunaria rediviva aes Frequency Rel o Row Turboveg No Column 1 Alnus glutinosa 4 Fig 55 Combined synoptic table with frequency and fidelity These tables can be exported into the current RTF export file as described in Section 1 9 60 Note In a combined synoptic table columns of relev s cannot be moved Other mouse functions are the same as in regular synoptic tables 1 11 10 Synoptic Table Export From the File menu select Export and Synoptic Table For more information see Section 1 9 4 1 11 11 Sorting Species in Synoptic Tables The Sort Species In Synoptic Table function is available from the Sorting menu or the Synoptic Table
72. c Table This type of synoptic table shows percentage constancy of species for each relev group This function considers only presence absence of the species without discrimination of cover values The value of percentage constancy is rounded to the nearest natural number in the interval from 0 to 100 except that all non zero values less than 1 are converted to the number 1 JUICE c documents and settingsMubom r tichy dokumenty 0_lubos juice kurz 10 3 2005Ydyje valley1 wct File Edit Species Sorting Synoptic Table Table Simulation Help Separator mJ X 4 Total time 3 days 12 h 49 min 34 sec Statistics u value hyp Percentage synoptic table Number of releves releves 202 Species 631 Poa trivialis Deschampsia cespitos Equisetum arvense Aegopodium podagrari Urtica dioica Geum urbanum Stellaria nemorum Galium aparine Pulmonaria obscura o Sambucus nigra Galeobdolon montanum Phalaris arundinacea Oxalis acetosella Carex sylvatica Glechoma hederacea h Carex remota Dryopteris filix mas Carpinus betulus Hepatica nobilis Dactylis polygama Quercus petraea agg Frequency 13 Rel TE Row 11 Poa trivialis Turboveg No Column 77 Start fe MALACKY Z number ro Gh obsah t04 SLOVNIK 1 oy juice Mier JUICE RE Gol PE 11 52 Fig 48 Synoptic table with percentage constancy 1 11 3 Categorical Synoptic Table Constancy can also be displayed in a catego
73. c tables it corresponds to relative frequency a 1 11 5 Fidelity Synoptic Table This table uses the fidelity concept as described in Section 1 10 The fidelity between the species and the relev group is displayed Highlighting thresholds are set in the Fidelity threshold section of the Synoptic tables tab of the Options menu See Section 1 11 1 Fidelity can be calculated either from presence absence or from quantitative cover data See Section 1 10 6 JUICE c documents and settingsWubomir tichy dokumenty 0_lubos juice kurz 10 3 2005Ydyje valley2 wct File Edi Species Sorting Synoptic Table Table Simulation Help Species Separator p x itae A 3 Total time 4 days 13 h 18 min 43 sec Statistics Phi costr Percentage synoptic table with fidelity index phi coefficient 8 columns Number of releves releves 202 Species 631 Poa trivialis Ranunculus repens Scirpus sylvaticus Caltha palustris Crepis paludosa Equisetum arvense Deschampsia cespitos Alnus glutinosa Lythrum salicaria Stellaria nemorum Aegopodium podagrari Phalaris arundinacea Lamium maculatum Sambucus nigra Alnus glutinosa Carex remota Carex sylvatica Dryopteris filix mas Avenella flexuosa Galium glaucum Inula ensifolia P PUlNOo oooN Frequency 16 Lythrum salicaria 6 Row Turboveg No Column Fig 51 Synoptic table with fidelity values Note Negative fidelities
74. ca Fidelity calculation 62 S Pinus sibirica 62 Cladonia rangiferina Add group into the table 66 Ptilium crista castrensis 69 Hylocomium splendens 69 Vaccinium vitis idaea 7 9 72 S Pinus sibirica Cover gt 72 Carex iljinii Header data analysis M E CERTI 100 Pleurozium schreberi 100 Polytrichum strictum Fig 73 Synoptic column calculated from all selected red relev s The list contains all species with cover value higher than 0 Species groups can be added directly into the table as a pseudospecies with before the group name Thus anything which can be done with species can be done with species groups Species groups and proper species can be combined into an aggregate group as explained in Section 2 3 6 2 3 5 Cocktail Algorithm for the Definition of Species Groups The COCKTAIL algorithm Bruelheide 1995 2000 was designed for statistical formation of sociological species groups It proceeds iteratively as follows Step 1 Construct an initial species group There are two ways to do so Step la Starting with preselected relev s typical of a known vegetation unit the algorithm begins by calculating all species fidelities to that vegetation unit and takes the species with the highest fidelity values as the starting species group Step 1b Start with a user defined species group based on the literature or previous analysis Step 2 The number of species of the species group is calculated for ea
75. ch have a similar distribution 8 Recolour these species and add them to the list in the COCKTAIL groups window as in step 3 above 9 Repeat beginning with step 4 until the group seems to be optimal for syntaxonomic classification This may be when it is similar to a group of diagnostic species traditionally recognized in the syntaxonomic literature An optimized group with relatively high fidelity values see below can be saved by writing its name into the combo box in the upper right corner and pressing the Add group into the table button Pressing the Synoptic col button displays the percentage synoptic column for the relev s of the selected colour Species Groups of Cocktail Classification hd Species list in selected group Synoptic column Polytrichum strictum 31 Anthoxanthum alpinum Pleurozium schreberi 31 Barbilophozia lycopodioides add f Refresh f Export Spaai Delete 31 Ledum palustre 31 Trientalis europaea Move species to the groups 31 Vaccinium uliginosum Species red 34 Sanionia uncinata 38 S Betula rotundifolia lt lt lt Mark in the table 38 Cladonia stellaris 41 S Cladonia arbuscula Delete species from the list 41 Bergenia crassifolia Min No Spec Rel Li Calamagrostis langsdorffii Observed Expected 41 Pohlia nutans Manual setting 45 Rhododendron aureum 55 Linnaea borealis Select relev s as red 55 Vaccinium myrtillus 62 S Lonicera altai
76. ch relev in the table The expected and observed cumulative distribution functions for relev s having 0 to k species are calculated The distributions intersection defines the required minimum number m of 81 species for a relev to belong to the vegetation unit The vegetation unit is defined by all relev s having m or more species belonging to the species group If there is no intersection between the observed and expected cumulative distribution then the algorithm aborts This is the case when species having fewer co occurrences than expected form the starting group Step 3 The occurrences of each species in the vegetation unit are counted and the fidelity is calculated Step 4 For each species in the species group fidelity value is tested against an initially fixed threshold fidelity If fidelity exceeds the threshold the algorithm proceeds to step 5 If not there are two possibilities Step 4a If this is the first iteration it is one of the initially selected species which does not exceed the threshold In this case the group is rejected and the algorithm aborts Step 4b If there has been at least one previous iteration it means the last species added caused another species s fidelity to decrease below the threshold The species with below threshold fidelity is removed and the algorithm does not try to add this species again until the group has been changed by adding a further species Step 5 All species not belongin
77. d be able to display the file in a fixed width font such as Courier Even though a TAB may appear the same as several spaces JUICE will interpret 33 it as a single character This could result in nonsense in the Species Data Column or it could make it impossible to tell JUICE where species names begin and end resulting in a blank Species Data Column Note The example window shown above illustrates that the external data file can contain several columns of data You can choose which column to import simply by specifying the range Note that if the column contains only one character the First character and Last character values will be the same If the option Connect selected data with all species is chosen the designated data will be imported into JUICE It is also possible to choose to import data only for species of a certain colour For example the option Mark all species with the value arc an abbreviation for archeophytes with the colour green can be used to colour species according to external data without actually importing the data It is also possible to only re colour species of a certain colour The example above will colour all species that have the abbreviation arc in the indicated column in the external data file 1 7 8 Species Group Tables In addition to the functions for writing data to the Species Data Column described above the Species Data submenu has four functions for d
78. d by species cover or abundance unless n and n are defined as in Section 1 10 6 The u value is defined as the deviation of the observed frequency of the species occurrence in the vegetation unit from the expected frequency compared with the standard deviation U hyp Cp I O yp Eq 5 The measure counts the number of standard deviations between the observed data and the result expected under the null hypothesis that species occurrences are independent of the target vegetation unit Thus ump provides a measure of the statistical significance of the observed relationship It can take on positive or negative values with the range depending on the number of relev s in the dataset The maximum u value vN 1 is achieved when N n np Fidelity rankings according to u will be identical to rankings according to chi squared or the phi coefficient b The phi coefficient of association between species and vegetation units Sokal amp Rohlf 1995 Chytry et al 2002 is a statistical measure of association between two categories which can be used as measure of fidelity It can be defined as Up ERE ee VN 1 nN N n N N Eq 6 46 Traditionally the phi coefficient considers only presence absence so fidelity values calculated using this coefficient are not influenced by species cover or abundance The values of the phi coefficient range from 1 to 1 but for convenience they are multiplied by 100 in the pro
79. dvanced Ordinal Scale Presence Absence Scale User Defined Scale Fig 16 Defined scales combo box in the Display parameters tab of the Options window Selecting User Defined Scale turns on the Modify button Clicking on this button opens the following window 21 User Defined Scale x nN Di t2 Ja pa Ok OM NN NN n2 t2 gt D emo m o D T NN eo J co nN we e c2 cen we J oO n3 ww Q2 T cn e Dan ceo M Ja n mn NN NN NN eo Nm N N N c2 4 x a m 4 S N N co n n O c 4 oak wh N NN NM NN NN ow Ja t en e T N NM NM NM NM NN NH e STSTSTSTSTSISEATS T7 pe o i eee E TL Md nN we ce P Di cn o y NN NM NM NM NM NN ow eo CISDPSTSTRTRTST ART BGasoeusessasgomsc NN NM NM HNN NK X repererer aay ayy or Di AT T TSTSTAISTSTSTAH STP TTT TT TT 81 D N os ow SPST ST OTST ISS STerererererereroro Terererererereroro 1 2 3 4 5 6 F 8 3 ow O OO Oi Urn NN NN NR NN NN NINININI NINININI n3 1 2 3 4 5 6 7 8 3 10 7 8 3 0 t n3 c2 D 13 p D cen em ceo co Braun Blanquet old scale Ordinal scale C Braun Blanquet new scale Presence absence Fig 17 Window for definition of user defined scale Every percentage number must be represented by one character which will be displayed in the table This scale is saved in the JUICE INI file
80. e Synoptic tables tab 63 Species exceeding these values will be displayed in bold style The sample species list in the lower part of the window will not change until the user presses the Preview button The user can export information for a single cluster column with the Export button The information for the entire data set can be exported by pressing the Export Clusters button All data are saved into the current RTF export file See Section 1 9 2 1 11 13 Uniqueness of Vegetation Unit Uniqueness was used in the paper of Chytry amp Tichy 2003 to identify unique vegetation units in the data set It expresses whether there are similar vegetation units of the same rank e g class or alliance A vegetation unit is unique if none of its diagnostic species has simultaneously diagnostic status in other vegetation units while its uniqueness decreases if it shares its diagnostic species with other vegetation units 1 11 13 1 Theoretical Background Uniqueness is calculated in two steps First an Asymmetric Similarity Index between every pair of vegetation units j and k is calculated as Diy Dik T AT Eq 8 i where is the fidelity of species i to vegetation unit j is the fidelity of species i to vegetation unit k and the sums only include species with gt 0 05 and gt 0 05 This index ranges in the interval 0 oo and yields the highest values for those pairs of vegetation units in
81. e C Documents and Settings Lubomir TichyDokumentyW0_lubos WJUICE Kubat KUBATS txt as er Use headers Append new header files Header data columns 1 Table number J Fig 3 Cornell condensed file import 3 and 4 step The next step is to specify the species list If the species list was loaded successfully and used for importing a previous table the program will automatically use this pre defined file as the source of full species names However it is possible to use a different species list or import species as abbreviations Header data will be loaded automatically from a pair of header data files with the same name as the CC file with suffixes EXP and STR It is also possible to import a table without header data or use a different header data file by clicking on the button Append new header files The program allows the use of headers from a different import package e g headers with a differently defined list of fields Such headers need not be complete It is recommended that you test their compatibility with the Test button The test will identify the number of relev s with headers Note Connecting table data with headers is only possible if the header data file contains the Relev number field with unique identification numbers 1 4 3 Species Lists Although the abbreviations used in Cornell condensed files should uniquely specify the species the final presentation of the table usually requires f
82. e N value can thus be used as a tool for modifying the properties of the phi coefficient with respect to weighting common or rare species The dependence of the phi coefficient on the relative size of the target vegetation unit can be tested in JUICE See Section 1 10 3 below and Section 1 11 on synoptic tables 1 10 3 Selecting Fidelity Measures and Standardization in JUICE The current fidelity measure is displayed on the Option Button on the left side of the Icon Bar See Section 1 5 1 Pressing this button opens the Fidelity measures tab of the Options window Options General External program paths Check List Import Fidelity measures Synoptic tables Display parameters Type of fidelity measure Different measures of fidelity are included in this Fidelity is calculated both for presence absence data program However we recommend to use phi coefficient without consideration of cover values and for all or hypergeometric u value For better understanding see cover data including zero values which are averaged the on line help and our paper Help for each table group phi coefficient Y Presence absence data Average cover u value hypergeometric adj phi coefficient Chi square Chi square adj G test likehood ratio G test likehood ratio adj u value binomial B Cc c Fig 43 Selecting a fidelity measure from the Fidelity measures tab of the Options window The user can s
83. e all relev s to be classified a single relev colour It is possible to classify every relev in the table by giving them all the same colour From the Analysis menu selecting the TWINSPAN function opens the following window of parameter settings Fig 62 TWINSPAN TWINSPAN parameters Pseudospecies cut levels gt a Values of cut l vels Minimum group size Maximum level of divisions Species sorting Make separators Run TWINSPAN Fig 62 JUICE window for TWINSPAN classification The value in the Pseudospecies cut levels field is the number of cut levels and the values in the Values of cut levels field are percentages corresponding to each cut level Both JUICE and TURBOVEG use whole percentage numbers from 1 to 95 for definition of cover scales Example In Fig 62 there are 5 cut levels of 0 2 5 10 and 20 A pseudospecies with a cover of 15 would exceed the fourth cut level but not the fifth It is possible to set a minimum group size and maximum number of divisions Note The maximum level of divisions displayed in JUICE is limited by the number of characters displayed in header data 6 so that a maximum of 64 groups can be displayed If the user wants a more detailed classification it is possible to classify the table in three steps 69 1 make an initial classification into 64 or fewer groups 2 mark part of the dataset with a unique relev colour
84. e clean and start over JUICE asks for confirmation before erasing the contents Another way to access this window is through the Export tab of the Options window clicking on the Change File button will open the export file name window 1 9 3 Table Export From the File menu select Export and Table This will open the Table Export window which has several options as can be seen in the picture below The File format box offers several different export formats The default MS WORD SDF appends the table to the RTF file described in the previous section All other formats prompt the user for a file name and export the data to this new file overwriting any previous file of the same name An MS EXCEL CD export will generate a stand alone CSV file accepted by spreadsheet programs See Section 1 4 4 This table is divided into two parts table and header An MS ACCESS export produces three files SPECIES TXT which contains species names layers and Ellenberg indicator values if they have been imported TABLE TXT with table data species number relev number and percentage cover and HEADER TXT with all header data These are only the default names the user can change them These files can be imported into a database program and connected as three tables A TEXT FILE export saves the file as simple text TXT See Section 1 4 5 A CSV HEAD TABLE create
85. e physically spliced together The standard RTF export MS WORD SDF file can contain separators There is also the option again only for RTF export to include only species which appear in at least 2 3 or 4 relev s listing the rarer species below the table in condensed form The length of the species name can be limited with a maximum length from 4 to 50 characters JUICE remembers original percentage cover values as whole numbers 1 95 but they are replaced by single characters during export except for the MS ACCESS format where percentage numbers are exported The list in the Table Export window contains a list of percentages Perc value on the left paired with the character to be exported Phyt code To change an export character double click on the corresponding percentage value Newly defined scales can be saved for future use by pressing the Save button Later they can be loaded with the Load button 1 9 4 Synoptic Table Export From the File menu select Export If JUICE is in Synoptic Table Display mode see Section 1 11 the Synoptic Table export function will be available This provides two possibilities to 1 Export the synoptic table into the current RTF export file for presentation or publication 2 Export the synoptic table into a Cornell condensed CC file for analysis by another program such as CANOCO ter Braak amp Smilauer 2002 or PC ORD McCune amp
86. e repainted 1 5 2 Mouse Functions Sorted by Similar Functionality Relev species selection eft Button lick noptic Table elect current species Separators Shift Left Button Pane Table Add remove species separator of selected Species ierarchy 15 Colours Right Button lick able and Header Repaint indicated relev with current relev able Relev olour epaint block of species with current species olour Click on the first species to be selected Manually moving species relev relev group Left Button lick and Drag Synoptic Table ove currently selected relev group Short Headers Editing species name or header data Left Button Double Click able and Header Open window with function for editing header able Header Data data 16 1 5 3 Colours As indicated in the previous section relev s and species can be assigned colours These colour codes can then be used for data processing analysis and classification Relev s and species each have eight colours available which can be selected from the Icon Bar The basic colour for relev s is white while for species it is black Colours enable organisation or analysis of data in a selected part of the data set Colour coding makes selection and manipulation of groups of relev s or species easier and quicker and sorted data may have a clearer structure The mouse commands for colouring relev s and species are described in Section 1 5 2 ab
87. ecies and Relev s i odo etre eae nete nr sa ttis 25 1 6 3 2 Other Species Sorting Functions sess eene enne etre 26 1 6 3 3 Other Relev Sorting Functions sss ener enemies 27 1 6 4 Autorepeat FunctioniJ s orici eo that baka a dei int es 29 1 7 Species Da ba ss cits c E M 30 1 7 1 VAAN CTS NC E n 30 1 7 2 Frequency and Cover Values eren edet tdt em nets 31 1 7 3 Sequence and Species Colour iidem etie neto m pee e aa es 31 1 7 4 Transformation of Species Data sse ener ener enne nnne 31 1 7 5 Statistics Summarizing Relev Data sess eee 31 1 7 6 Ellenberg Indicator Values ccccccccssessseescessessccesecesecsecaecseecaeeeaeeeaseeeesreseeeeseenaeeaeeaeceeeeenneeses 31 1 7 7 External Species Dafa 5 segete te eere peu ies oe ts 31 1 7 8 Species Group Tables essere ener nnne nnne trennen rre n eren enne 33 1 7 9 Species Data Averages Sis A euet D Ee bed e ie ed deett e dese Rita 34 1 7 10 Species Data Exports csetera eere tue es beds oes e so tts 34 1 8 Mitiags ripe RE 34 1 8 1 Identification Numbers 644 secs get aaa n ide ele Ait cone e ti aa a e a AN tens 34 1 8 2 ther Short Header Values io eet Lee tiet der deese ph RES 34 1 8 2 1 N mber of Specles a 5 et ene tu eet tuse I tut du EORR 35 1 8 2 2 Percentage GOVel 256 rete EEEE eee ter tr o PEPERIT gee et iu Se eer eur ee ERE Eee IR ses 35 1 8 2 3 Short
88. ed 44 1 9 9 Short Header Export eese eee eee eee et eee ei ee eee eie etae 44 1 10 The Fidelity Concept 45 1 10 1 Fidelity Measures cure cede tices ented aet e d ag VR e hehe ON Eee east eee 45 1 10 2 Fidelity Measurements for Vegetation Units of Unequal Size sess 48 1 10 3 Selecting Fidelity Measures and Standardization in JUICE eee 50 1 10 4 Fidelity Tests o oae teen tea addon ae ituiadiatie ee Saat 51 1 10 5 Tests of Data Structure Using Different Types of Standardization sss 52 1 10 6 Quantitative Fidelity Measures xh eee tee s ee e lei a de i rest 53 1 11 Symoptic Tables S orias SSeS 54 1 11 1 Synoptic Table Display 2 2x Re t eR e Be Ee ee ues 54 1 11 2 Percentage Constancy Synoptic Table sse ener 55 1 11 3 Categorical Synoptic Table ie eee ace wi Sh ie eee 55 1 11 4 Synoptic Table with Absolute Frequency Absolute Constancy ccccecseeseeesceseeeteeeteeeeenaees 56 1 11 5 Fidelity Synoptic Table nd e e e e e eet CDM D E Ce ae enge nee 57 1 11 6 Synoptic Tables and Cover Maximum Average Median Modus sess 57 1 11 7 Zlatnik S Combined Scale a aoro ate ee b oH OD abhi ee xc o e d 57 1 11 8 Average Cover Barkman s Total Cover Value nennen 58 1 11 9 Combined Synoptic Tables a caesa ed e Ue e ER DB ER Ue Orr CE E Pene eta 58 1 11 10 Synopt
89. elect one of thirteen different fidelity measures which are used to analyse presence absence or quantitative data in association with various program functions such as fidelity synoptic tables interspecific associations or the COCKTAIL method JUICE versions 6 3 57 and higher have relev group size standardization as discussed in Section 1 10 2 above The parameters of standardization can be adjusted in the Fidelity measures tab of the Options window 51 General External program paths Check List Import Fidelity measures Synoptic tables Display parameters Type of fidelity measure Different measures of fidelity are included in this Fidelity is calculated both for presence absence data program However we recommend to use phi coefficient without consideration of cover values and for all or hypergeometric u value For better understanding see cover data including zero values which are averaged the on line help and our paper Help for each table group phi coefficient hd Presence absence data Average cover Standardization of releve group size C A None B Only the size of target group Np N is fixed to C The size of All groups is standardised to C All but the last group equal size and Size of the target group 33 333 of the total dataset Calculate Fisher s exact test and give zero fidelity value to the species with significance Fig 44 Selecting a standardization method and Pre
90. els HOF are a hierarchical set of five models I flat with no response II monotone increasing HI monotone increasing to a plateau V symmetric unimodal and V asymmetric unimodal Four parameters are estimated using the non linear maximum likelihood estimation procedure described in Oksanen amp Minchin 2002b and further developed by Jari Oksanen in the gravy package However the gravy package gives ecologically unrealistic responses in some cases so some corrections have been made in this implementation As with any model response curves are just a simplification of reality and their shape is strongly dependent on the available dataset One major assumption which may be untrue is that the species has only one optimum along the gradient yielding unimodal response Bimodal response could have interesting and meaningful interpretation but it makes determination of optimum and species tolerance more complex and thus should be evaluated individually for particular species Therefore only unimodal or monotone response of species is considered here 2 8 1 5 How species optimum and tolerance are calculated The optimum is the value of gradient where the species has the highest probability of occurrence based on a particular model If the response curve is monotone decreasing or increasing the optimum is identical to the lowest or highest value of the gradient In the case of HOF models with flat segments of cu
91. ens amp Schamin e 2001 which is currently the most widespread database program for storing phytosociological data in Europe however it is also possible to import data into JUICE from a spreadsheet data format file In addition to basic functions useful for editing and publishing phytosociological tables the program includes various analysis functions such as Beals smoothing Ellenberg indicator values similarity indices beta diversity calculation interspecific associations and analysis of diagnostic dominant and constant species of synoptic tables and classification functions using COCKTAIL Bruelheide 1996 2001 TWINSPAN Hill 1979 or cluster analysis included in the PC ORD package McCune amp Mefford 1999 JUICE can create artificial data for testing Tables synoptic tables headers and different types of analysis including fidelities species groups indicator values and diagnostic species can be exported in four data formats 1 MS DOS text 2 Rich text format for word processors e g Microsoft WORD 3 spreadsheet format e g Microsoft EXCEL or 4 database format Microsoft ACCESS The program directly supports cooperation with the D MAP mapping package Morton 2005 and can create Cornell condensed files for other classification utilities such as CANOCO ter Braak amp Smilauer 2002 JUICE is continuously being developed since 1998 by the Working Group for Vegetation Science at the Department of Botany Ma
92. es more or fewer occurrences in the data set than the species already included in the species group This solution is recommended because groups of species with large differences in occurrence frequency would not be ecologically coherent their species might have roughly identical ecological optima but much more frequent species usually have broader ecological ranges After including the new species in the species group the group of relev s must be re defined and the fidelity coefficient of association between all species in the data set and the new group of relev s must be recalculated If this step causes the species group to disintegrate i e if some of the species not included in the species group now have a higher fidelity coefficient than some of the species included the group must be rejected By contrast if the species belonging to the group have the highest fidelity coefficients the group can be either accepted or further optimised by attempting to include additional species until all the remaining candidate species for inclusion either cause group disintegration or substantially change the ecological coherence of the group For a relev to be said to contain the species group it is not necessary for every species in the group to be present Bruelheide 1995 2000 defined the minimum required number of species of the group as the intersection of expected and observed cumulative distribution functions for relev s having 0 to k s
93. evel 10 characters space the code of the vegetation unit 5 characters space and the full name The second line gives the full Cocktail definition Selecting the Expert System function opens the following window Expert System Current expert file C Documents and Settings _ubomir TichyDokumentyW_lubosWUICE Expertni syst m 2004 expert_neles_2005 11 16_Basic txt CLASSIFIED RELEVE i4410 Create Expert File Part 4 esy 4 f Juncus inflexus Password wt 0 g Eriophorum latifolium z 93 34 2 88 Carex acuta Load ES File 33 3 4 E fH Cirsium rivulare A 50 0 4 A 8 Carex panicea nalpse r automatically 33 3 4 1 g Carex flacca se 25 0 4 s Filipendula vulgaris 4 5 0 4 8 Cratoneuron committat _ 24410 N fH Carex canescens Species name modif Modify Species Names Delete Juveniles Merge Same Spec Names Create Expert File Part 1 and 2 COCKTAIL CLASSIFICATION RBA0L Carici flavae Cratoneuretum Classify VIOLET relev s Copy to Clipboard Limit 0 100 Fig 75 The window for automatic classification of relev s with the expert system Two buttons allow the user to create Parts 1 2 and 4 automatically This requires the password esy04 The user must open the expert system source file with the Load ES File button The function is able to classify any newly imported data set The buttons Modify Species Names Delete Juveniles and Merge Same Spec Names en
94. ferent relev identification numbers e The Running Number of a relev tells what order it was in when the table was first imported This number is displayed by default Note This number is not constant When relev s are deleted from the table the running numbers of the remaining relev s are updated but the relative order of the running numbers remains unchanged For example if three relev s of lower running number are deleted relev s 44 and 45 will be re numbered 41 and 42 The name running number refers to the fact that this is the internal order in which JUICE processes or runs through the relev s e A relev s Turboveg Number is a constant unique relev identification number important for TURBOVEG users or other users who have their relev s identified by special numbers e When Group Number is selected every relev in the first group is given the number 1 every relev in the second group is given number 2 and so on Groups are defined by separators placed by the user See Section 1 5 4 If a relev is moved to a different group its group number will not change until this function is selected again e When Sequence Number is selected all relev s are re numbered in the order they currently appear If a relev is moved its sequence number will not be updated until this function is selected again 1 8 2 Other Short Header Values Short headers can contain up to six characters of inf
95. g the export from TURBOVEG Header data will import automatically with the CC file if all the files have the same name 3 All header data contained in an STR and EXP file can be imported separately From the File menu select Import and Header Data This is useful when a you wish to select different header data fields or b you need to add new header data to a spreadsheet format file or to any table without header data STR and EXP files are simple text files containing definitions and field values An STR file defines table data variable names and their starting character position in each line of the corresponding EXP file The first column defines the first character of the field the second column represents the name of the field Warning Each STR file must contain the field Relev number specifying where to find the relev s unique identification number An EXP file contains values of fields defined in a corresponding STR file e g relev information about the site and environmental factors All fields must be consistent with the format specified in the STR file 10 1 5 The Basics of Working with Tables This section describes the most basic functions for organising phytosociological data imported into JUICE The entire process is graphically oriented and more or less intuitive but the following text will explain some of the hidden features of the program 1 5 1 Table Window Components
96. g to the species group are sorted according to their fidelity value If any exceed the threshold fidelity the algorithm proceeds to step 6 If not the algorithm stops The species group is optimised when all species above the threshold are included Step 6 The species group is enlarged by including the single species with highest fidelity Iteration continues at step 2 When starting with preselected relev s belonging to a known syntaxon the vegetation unit is optimised in such a way that it is defined by differential species groups and the final composition of relev s in the group may be different than at the beginning Not all syntaxa can be defined by groups of differential species some are defined by dominance rather than by floristic composition Note Step 4a allows the user to assign a lower fidelity threshold making it more likely that the species group composition is not changed to such a degree that the initial species no longer have the highest fidelity This allows the formation of a number of species groups some with lower maximal fidelity than others Note that if the species in the initial group do not co occur with the vegetation type more than expected the group cannot be optimised 2 3 6 Definition of Relev Units by the Combination of Cocktail Groups The Group Aggregation function of the Analysis menu allows the user to define relev groups by combining presence of species groups and dominance of i
97. gram The highest phi value of 1 is achieved if the species occurs in all relev s of the vegetation unit and is absent elsewhere A positive value lower than 1 means that the species is absent from some relev s of the vegetation unit or present in some relev s outside the vegetation unit A value of 0 is obtained when the relative frequency of the species in the vegetation unit equals the relative frequency in the rest of the dataset thus indicating no relation between the target species and the target vegetation unit An advantage of the phi coefficient over some other statistical fidelity measures is its independence of data set size On the other hand the phi coefficient contains no information about statistical significance The phi coefficient along with u value and y is more or less dependent on the relative size of vegetation units If the target vegetation unit represents 10 of the total data set Fig 40a is relatively high for species that are not very common within the target unit provided the species are very rare outside the target unit However for species that are common outside the unit even the species with high frequency within the target vegetation unit are given low phi values On the other hand if the target unit makes up 50 of the total data set Fig 40b species must generally have higher frequencies within the target unit in order to have high phi values if a species has a high frequency within the target unit
98. h the standard central European check list Users from the Czech Republic can use the included KUBAT TXT check list with standard nomenclature Kub t et al 2002 A current version of this check list is available at http www sci muni cz botany juice jc05 che htm A new species list file can be defined in the Check List Import tab of the Options window available from the File menu If no species list is defined the program will open this window during the import of the CC file Options Fidelity measures Synoptic tables Display parameters Separators General External program paths Check List Import Check list file Cornell condensed files use species names as abbreviations The check list file allows to expand abbreviations into full species names automatically The check list file can be exported from TURBOVEG see JUICE on line help C Documents and Settings Lubomir Tichy Dokumenty 0 Iubos JUICE Kubat KUBATS txt File format Fields Open new check list file Fixed Length ID Number Abbrev Spec Name C Comma Delimited C Abbreviation Species Name Check list encoding T Short view of the check list file ID Humber Abbreviation Species Hame 50 IDN Sh cut Species name 1 ABIEALB Abies alba l2251 ABIEGRA bies grandis 2 ABIE SP Abies species ABITA A Abietinella abietina var abietina 4 5 ABITA H Abietinella abietina var hystricosa 3 ABITABI Abietinella abietina
99. he Analysis menu searches for optimal combinations of species that have similar distributions in the data set and can be used as sociological species groups in vegetation classification The reciprocal test of species associations is based on fidelity calculation A relev is considered to contain the species group if more than half of species of the group occur in it How to create a species group 1 Create an initial group of species using either the Interspecific Associations or the INI Groups function and give those species a unique colour 2 Select the Cocktail Groups function from the Analysis menu to open the window shown below 3 Add the starting group to the list box by choosing the correct colour and pressing the Add species button 79 Species Groups of Cocktail Classification Li Species list in selected group Polytrichum strictum Pleurozium schreberi Eo ETE Move species to the groups Species lt lt lt Mark in the table Delete species from the list Min No S IRel don ar E Observed Expected Select relev s as white Fidelity calculation Add group into the table Header data analysis aj Fig 71 The window for definition of Cocktail groups 4 The minimum number of species required for a relev to be said to contain the group Min No Spec Rel is calculated statistically so that the observed frequency of joint occurrence of several species
100. heir values in any column including the Max and Qual columns This sort must be within a single column of 87 the synoptic table selecting a colour which appears in multiple columns of the synoptic table i e in multiple relev groups as delimited by the separators produces an error message Pressing the Add value to short head writes the indicated value to the short headers of the selected relev s Sorting tj Sort relev s Sort by Freq Pos Fid FPFI Red Pos Fid PFDI Blue Frequency FAI Sea green Green Yellow Grey Ll Cancel Sort single column Fig 77 Sorting window in Matching function o o o o o o o Note Sorting across synoptic columns is not supported because such a sort will move relev s into different columns thus destroying the data on which the sort was based The expected use of this function is to sort a group single column of new relev s with respect to several groups of classified relev s The column to be sorted should be given a unique colour before switching to Similarity Index Display mode At present the colouring functions do not work in Similarity Index Display mode 2 4 2 3 Export of similarity index values The table with similarity index values may be exported by selecting from the Analysis menu Matching and Matching Export Text File The function is available only in Similarity Index Display mode Bug The
101. i coefficients in columns TAB1 and TAB2 min ED is the shortest distance between the column TAB and any of the columns TAB2 d TAB1 TAB2 Eq 10 and nr4g is the number of columns in the table TAB2 In this way we obtain the distance between the selected column of table TABI and the most similar column of table TAB2 divided by the average distance between the selected column and all other columns of TAB2 The distance between the tables TAB1 and TAB2 is computed by averaging the distance values for individual columns table of TABI Y d TAB1 TAB2 d TAB1 TAB2 i Eq 11 Arapi This distance is actually an asymmetric measure i e the distances d TAB1 TAB2 and d TAB2 TAB1 differ Therefore the same procedure is applied in the opposite direction min ED Aragi D Y ED min ED d TAB2 TABI Eq 12 gt d TAB2 TABI d TAB2 TABI Eq 13 Nr ape Finally the symmetric distance D TAB2 TABI between tables TAB2 and TABI is calculated as an average of d TAB1 TAB2 and d TAB2 TAB1 d TAB1 TAB2 d TAB2 TABI 5 This method can be used for calculating the symmetric distances between any two tables D TAB2 TABI Eq 14 1 11 16 2 Comparing Two Synoptic Tables within JUICE To use JUICE to compare two synoptic tables 1 Load the first table All records must be unique without duplicity of either species name or layer Such species must be merged
102. iagnostic constant and dominant species JUICE supports many different types of synoptic tables which can be used separately or in combination The columns of the synoptic tables are defined by the separators in the standard table Separators are placed or removed by holding down the Shift key and clicking on a relev with the left mouse button See Section 1 5 4 for more details To display a synoptic table select the type of table from the Synoptic Table menu To return to Standard Table Display select that type of table again 1 11 1 Synoptic Table Display The Synoptic Table Display has the same three parts as the Standard Table Display species short headers and synoptic table data see Section 1 5 1 Mouse functions in the synoptic table are described in Section 1 5 2 Several menu items are not available in Synoptic Table Display mode Note Columns in the synoptic table can be moved by clicking and dragging just as you do with relev s in Standard Table Display mode This gives a way to move a block of relev s without using colours Simply delimit the block with separators switch to a synoptic table move the column and switch back Values that exceed a certain threshold are highlighted These thresholds can be set in the Synoptic tables tab of the Options window It is also possible to select the highlighting colours as well as to turn the highlighting feature on or off Op tions Fidelity measures
103. ic Selection HOF Model HOF Model II HOF Model IlI HOF Model IV HOF Model Vv Gauss GLM GAM HOF GLM Models Automatic Selection Order of Polynom 1 Order of Polynom 2 Order of Polynom 3 GAM Models Automatic Selection Degrees of Freedom 3 Close R After the Process Degrees of Freedom 4 Degrees of Freedom 5 Fig 82 Window for setting parameters of the species response curves calculation With the R package JUICE can calculate 1 SRC for one selected species 2 SRC for comparison of several species or 3 SRC for all species When SRC is calculated for a single species several models can be displayed in one graph If multiple species are being compared the graph is displayed using only one selected SRC model Gaussian GLM GAM or HOF JUICE runs the R script automatically and the user must wait for results The results of calculating SRC for a single species are shown in Fig 83 Graph LJ Bergenia crassifolia Values I L L 1 L GAUS optimum 4 4007 GAUS min 3 70 GAUS max 5 3023 GAUS prob optimum 0 7107 o GAUS interval 1 6023 o GLM optimum 42733 5 GLH min 3 70 B GAM3 GLM max 5 4444 z o4 Y L GAUS BM pror optimum e GLM2 interval x Y P E HOEM amp LM model 2 a X m GAM optimum 3 70 o 02 7 E GAM min 3 70 a GAM max 5 3415 GAM test p 0 00 GAM interval 1 6415 GAM model 3 0
104. ic Table Export ecc cm e e ORE aoe nee ieee nee a ee Rete aes 60 1 11 11 Sorting Species in Synoptic Tables eene nennen 60 1 11 12 Analysis of Synoptic Columns Combining with Exporting the Results sess 61 1 11 13 Uniqueness of Vegetation Units 4 oec e DE TD DR ROTE EI Ue DER eee 63 1 11 13 1 Theoretical Backero nd 5 eere mte e rera ERE RR ge s 63 1 11 13 2 Uniquenessn JUICE eee ttt a e qu vid e e DUE ORE YU e DEN FR cane e 63 1 11 14 Average Values of Constancy Columns ssssssssssseeeseeeeeeenen eene ener ener nnne 64 1 11 15 Crispness of Classification cccccccesssesscescceseceseceseceeecscesaecseecaeeeaeeeaeeeeceeenseseaeeeaecaeeseceecneeeneeses 65 1 11 15 1 Theoretical Backeround 5 25 9 6 G0 d n nene Beate ime EOS 65 1 11 15 2 Optimal Use of This Function in JUICE sesesssseseeeeeeneenrer enne 65 1 11 16 Comparison of Two Synoptic Tables essere enemies 66 1 11 16 1 Theoretical B ckgro ndainin iiei mern erte naetiri roie eren enne 67 1 11 16 2 Comparing Two Synoptic Tables within JUICE eese 67 2 DATAANALYSIS 1st dieci e setciectsusntussutus aan dus uns ininda niea denne Tassen idni ndnani 68 2 1 PW INSPAN aE EE ENEE Vevs v eet os Vodka EY PE eub ong op VeRee tes ANO T RAENT AN 68 2 1 1 TWINSPAN in JUIGE nnno EP RI OR ao eta e PEE 68 2 1 2 Use of TWINSPAN as a Stand Alone Program eese 69 2 2 Cluster An
105. ices in JUICE Matching function Similarity indices may be used to classify relev s which are added to a data set already containing defined vegetation units 2 4 2 1 Procedure description Similarity indices are calculated using the Matching function of the Analysis menu and one of the three indices described in Section 2 4 1 This changes the table to Similarity Index Display mode To return to Standard Display mode make the same selection again Rows represent relev s and columns are the relev groups defined by separators The entries in the table are the similarity indices of the relev to the relev group i JUKE c documents and settings lubomir tichy dokumenty 0_lubos juice sajan sajan0 304 wet JA X Andyss Mel ia imm Tots tima 3 ups 13h 46 ce 7 ec CHEA A meri Frequency Positive Fidelity Index Humber of releves 2 Releves 311 Species 1028 Y w ON SWOOKVATSOONM OOH OCOMMOCOCSOMUOT EAS COMNOCOERSUsmTORCUSHOOOOCOCOCDOOUNZOUN CROCHHRENON CWO RSTUUUINSBOONDE CARBO HRS SCH UANGHH DOmOSCOOH CORONER ED H Re Be m Nerucowocoom 9 ja ME vesQgeoezueo POChaROCOOSCOON CA ONONAON HNO WOOO SC OCOO NOOMMVOCOCOCOCOO COC OUOSUUNSAOOHOOO OOOO NNN PPETI eevee N Oo4sANwWUwWuUHROCDOUWeHREHMUWUMOCHJADO CAD N m Ne B H m B NNE ow PES Be E HRS wOOBSUHPoOOSCOOeHENOD wee NH Ee NOORuUNSMMJOO OC C NOO CONONVOCONGOZIZZIOD M QU U Ui de NENNEN
106. ies Then from the Analysis menu select INI Groups This opens the window shown in Fig 70 The currently selected species or the group of coloured species can be added to the list on the right by pressing the appropriate button When the RUN the analysis button is pressed the function calculates interspecific associations of the first species in the list with all the other species and adds the most closely associated species to the list The number of species considered for addition corresponds to the number in the Nr of species to be added box The procedure is repeated for each species in the list including those which are newly added Note that this process terminates because eventually every most closely associated species will already be in the list The Mark in the table button causes every species in the list to be assigned the selected colour back in the table 78 Initiation of Groups EJ Species a Polytrichum strictum red Pleurozium schreberi lt lt lt Mark in the table Delete species from the list Nr of species to be added za RUN the analysis Fig 70 The window for creating an initial group of species Warning A high number of species to be added in larger tables can cause a very long cycle which cannot be interrupted without terminating JUICE The user is advised to save all work beforehand 2 3 4 Cocktail Groups Definition The function Cocktail Groups in t
107. ification Classification methods usually produce results with hierarchically distributed clusters Sometimes it is difficult to determine an optimal number of clusters providing the highest separation power for species JUICE includes a function proposed in the paper of Botta Dukat et al 2005 for identifying the optimal number of clusters 1 11 15 1 Theoretical Background The calculation is based on G statistics which can be easily calculated for contingency tables of any size Sokal amp Rohlf 1995 Calculated for a 2 x c contingency table where c is the number of clusters the G statistic does not measure the fidelity of a species to individual clusters but the species s capacity to distinguish the clusters within a given partition the separation power of the species The average of separation powers is called the crispness of the classification The higher the average separation power the better the clusters can be distinguished by the diagnostic species i e the better the classification The problem is that the expected separation power and consequently the crispness of classification increase with increasing number of clusters even if relev s are assigned randomly to clusters This effect has to be eliminated before comparing partitions with different numbers of clusters In a random classification separation powers have approximately a chi squared distribution with c 1 degrees of freedom Sokal amp Rohlf 1
108. ing on the first species and Shift clicking on the last species Undelete Species Hidden species Sele s to undelete Abietinella abietina var hystricosa Acer campestre Acer campestre Acer species Aconitum callibotryon Fig 21 Window for undeleting species Relev s can also be deleted From the Relev menu select Delete colour Relev s However there is no way to restore a deleted relev 1 6 3 Sorting Species Species Data and Relev s Several types of sorting are available from the Sorting menu In addition to basic species and relev sorts JUICE can also sort according to headers average Ellenberg indicator values clusters calculated in PC ORD etc 1 6 3 1 Sorting Species and Relev s From the Sorting menu choose either Sort Species Ctrl D or Sort Relev s Ctrl U These functions sort species by relev s or relev s by species The sorting hierarchy is according to 1 frequency 2 order of relev s species and 3 cover This means that a species which occurs in more relev s is ranked higher If two species occur in the same number of relev s the one that occurs in the relev listed first in the table is ranked higher Two species that have the same frequency and the same first relev are ranked according to cover An analogous system applies to relev s Colours can be used to limit the list of species and relev s to be sorted o o o o o o o
109. ion makes smaller cover values of less abundant species comparable with higher cover values of very abundant species The user specifies the number of clusters This number does not influence the time required to perform the calculations The user selects one of 7 distance measures resemblances and one of 8 group linkage methods 2 2 1 3 Results and Cluster Tree Cluster analysis begins when the Continue gt gt gt button is pressed The classification process is managed externally by sending key codes which makes it sensitive to any use of the keyboard The message window with this information remains visible while PC ORD is running When the classification is finished the table is sorted in the calculated order and the following results window Fig 65 is displayed 71 PC ORD cluster analysis E Please Select No of Divisions IET M Eee qo eth Help Tree Cancel Create Clusters Fig 65 Result of PC ORD cluster analysis The window displays a list of group numbers Selecting a group and pressing the Create Clusters button or double clicking on a group creates clusters Clusters will be divided in the table by separators Below the list are check boxes for copying cluster numbers into the short headers and for closing the window after selection of the cluster number If the short headers are being used for other information environmental variable etc switch off the Copy cluste
110. ion of diagnostic species by measuring fidelity proposed by Chytry et al 2002 is included in JUICE The program has 13 different fidelity measures available from the Fidelity measures tab of the Options window See Section 1 10 3 They are for use in synoptic tables as described in Section 1 11 1 10 1 Fidelity Measures The fidelity measures available in JUICE can consider either binary presence absence data or quantitative cover data For simplicity only four fidelity measures using presence absence data will be discussed in this section Measures that take cover into account will be discussed in Section 1 10 6 For details on other fidelity measures refer to Chytry et al 2002 Here the same notation is used as in Bruelheide 1995 2000 and Chytry et al 2002 N number of relev s in the data set N number of relev s in the target vegetation unit n number of occurrences of the species in the data set Np number of occurrences of the species in the target vegetation unit To see how these quantities are defined when quantitative cover data are taken into account see Section 1 10 6 a The u value for a hypergeometric distribution Chytry et al 2002 further referred to as Unyp compares the observed number of occurrences of the species in the vegetation unit np with the expected number of occurrences u n N N The fidelity values calculated using this coefficient are not influence
111. ions separators etc are reversible Often an incorrect step can be repaired with the Undo function in the Edit menu Notes Undo is only supported for one operation It is not possible to Undo multiple mistakes In addition the Undo function does not support changes in data structure species names cover codes deleting relev s and species etc We recommend that you create backup WCT files as frequently as possible 24 1 6 Editing Tables Analysis classification or publication of phytosociological tables requires a clear data set without residuals or incorrectly identified species and with correct nomenclatorial background If the source data set consists of relev s in different scales and taxonomical concepts sampled by different authors it is necessary to unify the data This section describes how to accomplish these operations 1 6 1 Merging Species Species should be merged immediately after importing the table To merge nominally different taxa into one taxon follow these steps 1 Sort the species list into alphabetical order From the Sorting menu select Sort species alphabetically and ALL 2 Choose a secondary species colour Species of this colour will be merged To select the colour hold down the Ctrl key and click on the colour on the Icon Bar The box labelled lt Ctrl gt will display the selected colour 3 Mark species to be aggregated with this secondary co
112. isplaying species group tables Species group tables are similar to synoptic tables but constancy is defined not for groups of relev s but for groups of species Each constancy row is identified by the name of the first species in the group The species data is displayed if it is the same for all species in the group otherwise the field is filled with repetitions of the letter X Each constancy row has 6 characters Digits are displayed vertically The species group table can display frequency percentage values 0 100 categories I V total cover aggregate cover values from 0 to 100 are calculated as described in Section 1 6 1 or absolute species numbers To return to Standard Display select Species Group Table again JUICE c documents and settingsWubom r tichy dokumenty 0_lubos juice crambe2 wct File Edit Species Relev s Head Sorting Separators Synoptic Tal Indicator Values Analysis Table Simulation Help saa Reeve white v EK ies r ssi x Statistics Phi costr Total time 3 days 0 h 57 min 41 sec Running number Releves 77 6676 6 7777777 6643333333555212433344 Species 360 586613542701452347033210546678857278915 GROUP Onobrychis viciifoli XXX 1 11 1 1 11 2111111111 Total 141 046496006666466430841991199417111021765 GROUP Orchis militaris ag Ball Lo a LG Total 18 060666660670060011000600100161077006000 GROUP Stipa pennata 111111111 1 11111221111112111122111 Total 168 26612544588561
113. lassification techniques not available in major statistical packages PC ORD cluster analysis can be accomplished directly from JUICE Other PC ORD analyses must be done manually using exported files See Section 1 9 3 A demo version of PC ORD and information about ordering is available at http home centurytel net mjm pcordwin htm 2 2 1 1 Initiation of the Connection Before using PC ORD with JUICE PC ORD must be installed on the computer and the location of the PC ORD program file must be specified in the External Program Paths tab of the Options window JUICE will use this path for temporary files and will run the program automatically Fidelity Measures Synoptic Tables Display Parameters General Export External Program Paths Check List Import External programs called directly from JUICE The JUICE program allows a co operation with several classification and mapping programs distributed by other authors Please fill correct path of these programs TWINSPAN si D MAP Change file PC ORD Change file MULVA Change file SYN TAX 2000 Change file R PROJECT Change file Fig 63 The External Program Paths tab of the Options window 70 2 2 1 2 Automation of Cluster Analysis Within JUICE Cluster analysis in JUICE is available from the menu Analysis and Cluster Analysis The following window Fig 64 appears Cluster Analysis PC ORD 4 1
114. ld Select header field Table number Author code Altitude m Aspect degrees Slope degrees Cover total Cover tree layer Cover shrub layer Cover herb layer Cover moss layer Sort Separators Continue gt gt gt Add first characters to the short head Char lenght Cancel Fig 25 Sorting relev s by selected header data Sort Relev s by Clusters is used to display clusters computed by another program either PC ORD or MULVA With PC ORD the cluster information must be saved in a comma delimited file called MATRIX2 CSV This file can be imported with the function Sort Relev s By Clusters PC ORD With MULVA the table should be exported as a MULVA input file from the File menu choose Export and Mulva Input File Once analysed in MULVA the resulting file PRINDA without suffix can be loaded with the function Sort Relev s By Clusters MULVA The sorting window using PC ORD outputs is shown here PC ORD cluster analysis Please select No of divisions ister No o ean E Help Tree Cancel Create clusters Fig 26 Sorting species by PC ORD clusters 29 Double clicking on the desired level of classification will sort the data set accordingly The Tree button can be used to display a simple clustering hierarchy The Y axis is not scaled in this chart JERA v ims I Total time ug Hee ES Statistic
115. lla vulda pe eh ea ee 2k Se nhe ER eR ES Alisma aramineum s seca coterie A oe E EE e Seeded Bo neon ed Alisma lanceolatum Nop As LL meaner Lua rc Ere Alisma plantaqo aqua FEE c E E E AA A Allium anqulosum BENED Allium carinatum E EEUU TEIEEEEECDI c eves AES Allium oleraceum CETT Mn et ES Arp Pgh CHR re Fa rag ed aR ta cs ESAS pak AUD ERR Oy PEAT SLOT RENE ES ENE Allium scorodoprasum Hin cis eu e 8 EGO RS So Ce du dS ORAS COQUE Allium vineale C A E E SE EE c a aide Hate Pepe ea Pes Frequency 591 Rel o Row 22 Turboveg No Column 26 3 S Anthoxanthum odoratum 46699 7 Bam J Start Basta mam we m os 42v aso En 100 Fig 5 Main window of the program The table window is divided into three parts short headers species names and table data The Species Data Column can hold additional information about a species such as layer biological information or Ellenberg indicator value Ellenberg et al 1992 The Menu Bar with twelve items is at the top of the window The Icon Bar is below the Menu Bar Some menu functions are directly accessible as icons See the figure below Open file Header data Move relev s right Move species down Sort species Save file Find species relev Sort relev s Move species up Merge species Increase text Layer view Reset relev colour 2nd spec colour Divide species Decrease text Relev colour Move rel left Species colour Reset spec colour Separators gne
116. lour Hold down the Ctrl key while right clicking on the species name Make sure no other species are marked with this colour 4 From the Species menu select Merge selected Species or press Ctrl L 5 Confirm name and layer of the new aggregated species All the species contained in the resulting aggregated species are automatically deleted from the data set The cover of the aggregation is calculated under the assumption that covers can overlap and that they do so independently of each other Example Species 1 cover 50 Species 2 cover 30 Species 3 cover 20 JUICE starts with 50 the cover of the first species Under the independence assumption the second species covers 30 of the area covered by Species 1 and 30 of the area not covered by Species 1 This gives an additional 15 cover 0 5 x 0 3 0 15 These two species occupy together 0 50 0 15 0 65 65 of the sample plot Species 3 covers 20 of this covered area and 20 of the remaining 35 The area covered only by Species 3 is 7 0 20 x 0 35 0 07 The total area covered by at least one of the three species will be 0 50 1 00 0 50 x 0 30 1 00 0 65 x 0 20 0 72 72 94 Note that this is equivalent to calculating the area not covered by any species Under the independence assumption this area is 1 0 50 x 1 0 30 x 1 0 20 0 28 28 This means 72 is covered by at least one of the three species This independe
117. mat An example of a simple text file denoting ploidy level accepted by JUICE is show below 32 Aira caryophyllea Aira praecox Aira species Ajuga chamaepitys Ajuga genevensis Ajuga pyramidalis Ajuga reptans Ajuga species Alcea biennis Alcea rosea Alcea species Alchemilla alpina D I Pb eB ITN D co M I Each line has the same number of characters Species name and other information occupy the same position From the Species menu select Species Data and External Species Data Enter the name of the file containing the external data The following window will appear Species Data From External File mmus E PE 4 ce xum pem TE E a re TET ems E em Oxybaphus nyctagineus naturalised t accidental deliberate botan us Tetragonia tetragonoides casual l deliberate food Portulaca oleracea subsp oleracea naturalised pi 5 accidental none Portulaca grandiflora casual l deliberate ornam Calandrinia compressa casual 1 deliberate ornam Claytonia perfoliata casual l deliberate food Claytonia alsinoides naturalized 1 deliberate ornam Quercus r bra e ea S deliberate mood landst mam Almas rugosa naturalised 2 deliberate ornamental Corylus maxima casual f deliberate ornamental food Juglans regia naturalised 5 deliberate food wood Phytolacca americana casual f deliberate dye medic Phytolacca esculenta naturalised 3 deliberate dye medic Mirabilis jalapa casual 2 delibera
118. menu It is only active during Synoptic Table Display mode Selecting this function opens the following window Sort by Sort Species Relative frequency Fidelity measure Maximum cover Black Average cover Red Cover median Cover modus Blue Ziatnik s value Sea green Total cover ratio Wolfgang s measure Green Yellow Sort Violet Entire dataset MI Entire dataset without rey last column Single column o o o o o ie o o All Column Cancel Fig 56 The synoptic table Sorting window Species can be sorted according to several different measures The user has the option to sort the Entire dataset the Entire dataset without last column or a Single column The window contains a slide bar for selecting which column will be sorted by the Single Column sort The function can be restricted to sort only species of a certain colour The Entire dataset sort is a cumulative sorting function 1 When this function is called the user is prompted to define a sorting cut off level The default cut off level for relative frequency and fidelity will be the same as the Lower threshold defined in the Synoptic tables tab of the Options window See Section 1 11 1 Furthermore changing this cut off level for sorting changes the Lower threshold for highlighting 2 Once the cut off level is defined the first column is sorted 3 Species
119. nce assumption is most appropriate when merging different layers for the same species When merging distinct species of one layer into an aggregate it may be more reasonable to assume the covers are mutually exclusive In this case the average cover values computed by JUICE may underestimate the resulting cover however other aggregate statistics such as presence absence remain valid Note 1 The default name and layer of the aggregation correspond to the first species in the list If you wish to use a different species to provide the default name and layer simply move that species to the top of the list This information can also be entered manually 25 Note 2 Merged species are removed from the table However each merged species can be returned to the table with the function Undelete Species from the Species menu See Section 1 6 2 1 6 2 Deleting and Undeleting Species and Relev s To delete species from the table give them a colour and select Delete lt colour gt Species from the Species menu Species which have been deleted from the table or merged into an aggregation can be retrieved using the Undelete Species function from the Species menu This opens a list of deleted species which can be sorted by layer name or time of deletion Select the species to be restored and press the Undelete button Multiple species can be selected with Ctrl click A block of species can be selected by click
120. ndividual species It uses species groups created in the table see Sections 2 3 4 and 2 3 5 and dominant species which are defined by cover values exceeding a selected threshold Species groups and dominant species are combined using the logical operators AND OR and NOT which functions as AND NOT with the hierarchy defined by parentheses 82 Relev Selection Relev selection O White Acinos arven mr pa Aconitum plicatum Red Adonis aestivalis Blue Aira praecox Alliaria petiolata Amaranthus retroflexus Andromeda polifolia Anthoxanthum odoratum O Sea green fe Aphanes arvensis O Yellow o o Green Arabidopsis thaliana s Arabis auriculata Violet Armeria serpentini P Arnoseris minima rey Li Li J lt Ulmus glabra gt OR lt Carpinus betulusUP25 gt AND lt Mercurialis perennis OR lt Urtica dioica gt oct wor fanof orfejoj e p Dr make setection fi E Show definition Fig 74 The window for selection of relev s containing a combination of species groups and dominants Species group names created by the Cocktail analysis function begin with the characters Names of dominant species are not preceded by characters but have suffixes such as UP05 or UP25 For example UP05 means that the species is considered if its cover in the given relev is higher than 5 UP25 means higher than 25 96 A sample query is given below lt Ulmus glabra gt OR lt Carpinus betulusUP25 gt
121. ns S Veronica chamaedrys agg S Ranunculus acris Rte tet t PRR RBBB RN NOR Table number 6839 Relev number 587205 Biblioreference 587100 No table in publ No relev in table Year Month Day Author code Syntaxon code Relev area m2 283 Hydrocotyle vulgaris 1 Rel o Turboveg No Frequency Fig 12 Selected relev displayed at the right site of the window 1 5 7 Editing Species and Header Data JUICE is not database oriented software for data archiving All changes are saved only in the current table without impacting a database source Therefore we strongly recommend archiving phytosociological or ecological data in a database program e g TURBOVEG and making all changes in the original data set 19 However it is possible to modify some data in an existing JUICE file 1 5 8 JUICE New species name Impatiens parviflora Species data Fig 13 The window for editing species names layers and species data a Species names layers and species data values can be edited by double clicking on the species name b Cover values displayed in the table can be edited from the Table Simulation menu Selecting Add Remove Species Cover opens a window for specifying the value to be written into the table Warning In cover value editing mode the program changes mouse functions The left mouse button writes the specified cover value to the indic
122. o include all species Data can be sorted in alphabetical or numerical order and the order can be ascending or descending Species Sorting Parameters Sort species data o Black Sort species data Alphabetically Numerically Close Continue gt gt gt Fig 23 Window with species sorting parameters Note Numbers should be sorted in numerical order 1 2 3 11 12 13 21 22 23 while text strings must be sorted in alphabetical order If the above numbers are sorted in alphabetical order the result is 1 11 12 13 2 21 22 23 3 Sort Species in Synoptic Table is available only in Synoptic Table Display mode The Sorting window has several options Species can be sorted by various criteria such as constancy fidelity or average cover Colour selection can be used to restrict the species included in the sort This function is described in more detail in Section 1 11 11 Sort species data In ascending order In descending order All 27 Sort Species by Median Cover sorts species by the median of all non zero cover values in the table In relatively large and diverse data sets this can be used for deducing the theoretical abundance of each species under optimal conditions The median cover value is not displayed after sorting To view the median cover value and other species information select Species Statistics from the Species menu To return to Standard Display mode select
123. of reserving 6 characters for header data you can combine several header data values for one relev For example it is possible to use one file to import class number 2 digits into the first two characters and another file to import year 4 digits into the last four characters of the short header field This will enable you to sort the table with classes as the main criterion and year as the secondary criterion 39 1 8 3 Colouring Relev s According to Short Header Relev s do not have to be coloured manually It is possible to colour them according to the data in the short headers From the Head menu select Short Header Selection This opens the window shown below Choose a colour enter a value and specify whether to mark headers greater than less than or equal to the entered value Short Header Selection white X headers ZEE Numerically Parameters lt nens ok Fig 35 Short header selection window Note Alphabetical comparison should be used with text and numerical with numbers The example below shows what happens when numbers are compared alphabetically Numerical comparison 12 345 gt 9 8765 Alphabetical comparison 12 345 9 8765 With numerical comparison text is interpreted as a number If the text contains no digits it has a value of zero 1 8 4 Short Header Averages Minima and Maxima JUICE can measure the average minimum or maximum value of the short headers
124. of the R software package It must be downloaded from http cc oulu fi 7Ejarioksa softhelp softalist html Select the latest version of the gravy ZIP file in the ZIP files for Windows folder Download and unzip into the folder c Program Files R R 2 2 1 library To check if installation of both R and gravy packages was successful run the R program and into the command line write library gravy An error message indicates that the gravy package was not installed properly The R script is the file juicel txt from the web page http botanika bf jcu cz david hof php Download this file into the folder c Program Files R R 2 2 1 bin This file enables JUICE to communicate with the R package 2 8 3 Calculation of Species Response Curves SRC in JUICE JUICE is prepared for SRC calculation after the installation of the components mentioned in Section 2 8 2 Calculation is based on a single environmental gradient which must be saved into Short Headers If the user has e g pH values in header data it is necessary to copy them into the Short Headers field as described in Section 1 8 2 3 Then from the Analysis menu select Species Response Curves This opens the following window 94 Gradient analysis Name of the Gradient in Short Headers Field Transformation of cover values Q Presence Absence Square Root Transformation Percentage Val No Transt HOF Models Automat
125. ogens biomass productivity KUBAT TXT species check list useful for phytosociologists from the Czech Republic The check list has a correct nomenclature published in Kub t et al 2002 NEWFLORA TXT older species check list acceptable in the Czech Republic Slovakia Austria and Hungary based on unpublished nomenclature Users from other countries should export the check list from TURBOVEG See Section 1 4 2 TWINSPAN EXE a modified version of the famous freely distributed classification program This version is integrated with JUICE It can also be run as a stand alone DOS program At the beginning of the installation the user will have to specify if the program should check for older versions and reinstall them This operation is strongly recommended but in most cases it is possible to have several versions of JUICE installed on one computer When installation is complete you may run the program from the icon group JUICE 6 3 Note 1 The latest version of the program can be also downloaded as a simple EXE file which must be copied directly into the existing JUICE directory usually C Program FilesJUICE 6 3 Update your program frequently If you install the program on a new computer please use the full installation Note 2 In a network we recommend installing JUICE on each computer individually Otherwise there might be conflicts between multiple copies trying to access the same directory on the host computer The
126. orm to clipboard Shift Left Button eae ake remove separator line to the right of currentl selected relev trl Right Button D ee currently selected species with curren secondary species colour When the cursor is in the table data MEER ouble Click Display list of species in selected relev and save selected relev in text form to clipboard trl Right Button nee currently selected species with curren secondary species colour In synoptic tables gt JUICE c juice_fin 1 louky_ass_str_del01 wct File Edit Species Sorting Synoptic Table Table Simulation Help Percentage synoptic table Number of releves releves 1210 Species 606 Achillea aspleniifol S Achillea millefo Achillea ptarmica Acinos arvensis S Aconitum napelus Adoxa moschatellina Aegopodium podagrari S Agrimonia eupato Agropyron caninum S Agrostis canina Agrostis capillaris S Agrostis stoloni Ajuga genevensis Ajuga reptans S Alchemilla hybri S Alchemilla vulga Alisma gramineum Alisma lanceolatum Alisma plantago aqua Allium angulosum Allium carinatum Allium oleraceum Allium scorodoprasum Allium vineale we AX Total time 2 days 6 h 22 min 47 sec 1 Achillea aspleniifolia Frequency Rel flo Turboveg No Fig 7 Synoptic table Functions are slightly different in synoptic tables When the cursor is in the short headers Left Button Click and Drag When
127. ormat C Documents and Settings _ubomir TichyDokumenty 0_lubosWUICE kropac2 cc File parameters Species 432 Zero elements 92 2 Relev s 712 File size 357 257 kB 1 000 2 000 will b value 3 000 will b value 000 will r value 13 000 will 23 000 will b 9 000 will be s 2 000 will 000 will 88 000 will be saved Fig 2 Cornell condensed file import 1 and 2 step The basic statistics of the selected file are presented under File parameters If they are incorrect this indicates that the file is not suitable for this type of import TURBOVEG files are defined with species abbreviations in the format 7 1 7 characters encode the species name and the last character is reserved for the layer number however CC files without layer identification can be imported by selecting the option 8 characters The scale is detected automatically but it can also be selected manually TURBOVEG exports only percentage values The program checks all cover values and tries to convert them to whole numbers from 1 to 95 If a value cannot be imported automatically it is necessary to supply the correct conversion Double clicking on a line in the Covers list box opens a window for entering the corresponding value Import Manager For TURBOVEG Or Cornell Condensed File Step 4 5 Table file Species 709 Zero elements 96 1 Relev s 600 File size 254 209 kB Check list Use check Append new ch
128. ormation about the relev They can be used to distinguish relev groups in the data set and define them with separators or colours as described in Section 1 8 3 The functions described below are found in the Head menu under Store Values To Short Headers 35 1 8 2 1 Number of Species The function Count colour Species counts all species of the selected colour which could correspond to all mosses or trees for example Species with the same name originally recorded in several different layers are virtually merged and counted only once 1 8 2 2 Percentage Cover The function Percentage Abundance Of colour Species calculates total percentage cover of the selected species This can be used to distinguish relev s in which the selected group of species is dominant from those relev s in which the group plays a marginal role The total cover is calculated on the assumption that species covers can overlap and that they do so independently See Section 1 6 1 for an example 1 8 2 3 Short Headers with Header Data The Header Data function allows the user to write header data to the short headers Any header data can be stored in the short headers however because the short headers can not display more than six characters some header data fields such as locality geology or remarks are not appropriate for conversion Move header field to short headers Select header field Author code Altitude m As
129. ove Colours can also be assigned according to information in the short headers as described in Section 1 8 3 Using colour coding to gather relev s and species into blocks is explained in Section 1 5 5 below 1 5 4 Separators Separators divide a table into sections This is necessary for defining synoptic tables or species group tables Such tables are used to analyse constancy fidelity and similar measures of a relev group s relation to species or a species group s relation to relev s Separators are placed or removed by holding down the Shift key and clicking on the relev or species with the left mouse button See Section 1 5 2 The separator appears on the right side of the indicated relev or below the indicated species Separator zu ienie h IX 1 2 3 4 5 6 Fig 9 Separator switches in the Menu bar The program includes the option to work with six hierarchical layers of separators Level 1 can be used for major groupings level 2 can be used for subgroupings and so on When the hierarchy is turned on the user can choose how many levels will be displayed The top level level 1 is always on while the bottom level level 6 is only on if the user chooses to display all 6 levels These switches can be found on the Icon Bar or under the Separators tab of the Options window which can be opened from the File menu Note Moving synoptic columns see Section 1 5 2 will destroy relev separator hiera
130. pecies k being the number of all species included in the group However our studies showed that this criterion strongly depends on the data set structure and tends to yield a low minimum number of species if the group consists of species that are rare in the data set and a high minimum number of species if the group mainly includes common species Therefore we propose a simpler criterion taking half of the species of the group as the minimum number e g at least 2 of 4 or 3 of 5 After defining several species groups the Cocktail method creates definitions of vegetation units by combinations of species groups using logical operators such as AND OR and AND NOT Bruelheide 1997 For example a relev is assigned to vegetation type X if it contains species group A and at least one of the species groups B or C and at the same time the species group D is absent 76 X A AND B OR C AND NOT D Note Species groups may be combined with single species in the same way if the user uses the concepts of both dominants and species groups 2 3 1 Co Occurring Species To analyse co occurring species the user selects a species by clicking on the species in the table and from the Analysis menu calls the function Co occurring Species This calculates the frequency of each species in the relev s containing the selected species and opens the Co occurring Species window Fig 68 Co occurring Species Circaea alpina
131. pecies data makes the interpretation of the resulting response curve more clear Information about species abundance or cover respectively is affected by complex factors including competitive relations species morphology and other biotic aspects which are difficult to interpret especially when taken in combination The presence absence transformation reduces the influence of these factors Austin 2002 2 8 1 2 Technical Notes on Particular Modelling Strategies 1 The bell shaped response curve is not based on the classical Gaussian equation but on a simplified equivalent polynomial model ter Braak amp Looman 1986 Oksanen amp Minchin 2002a which can be easily fit using generalized linear models with logistic link function for presence absence data and gives results close to a true Gaussian curve 2 Generalized linear models GLM are included in the standard R package Available models are linear quadratic and cubic polynomials of degree 1 2 and 3 respectively as polynomials of higher degree are not always unimodal The logit link function is used Selection of model 92 can be done manually or automatically based on AIC test criteria selecting the model with the lowest deviance of data 3 General additive models GAM are included in the mgcv library included in the R package Models with 3 4 or 5 degrees of freedom are available Automatic selection is based on AIC test criteria 4 Huisman Olff Fresco mod
132. pect degrees Slope degrees cance Continue gt gt gt Fig 31 Moving first part of selected header field into sort headers 1 8 2 4 Shannon Wiener Index and Evenness The functions Shannon Wiener Index and Equitability write these values to the short headers The Shannon Wiener index is one of several measurements of biodiversity Hill 1973 It takes into account the number of species and the evenness of the species The index is increased either by having more unique species or by having a greater evenness S H p inp gt Eq 1 i l where S is the number of species and pi is the proportion of the individual species cover relative to the total cover The program calculates two different measures of evenness Shannon s equitability Ex proposed by Pielou 1975 is calculated by dividing H by H ma Here H max InS Equitability has a value between O and 1 with 1 representing complete evenness BE H H ax H flus Eq 2 36 This value is also displayed in the Relev Overview Window see Section 1 5 6 which can be opened by double clicking on the short header The second measure of evenness available in JUICE is Evar which is equivalent to the arctan transformed Gaussian width Smith amp Wilson 1996 hy F o In p j 2 5 n p TE Eq 3 E a 1 arctan Z amp L m S 1 8 2 5 Geographical Position Index The Geogr
133. r No check box Note This window is always on top enabling it to be used in combination with synoptic tables sorts of species in synoptic tables crispness calculations etc which makes cluster analysis much more useful and clear The advantage of this approach is clearly demonstrated by a small example Display a classified table as a synoptic table with fidelities e g select fidelity coefficient Phi and double click on Gr_2 in the window Sort the entire synoptic table by a selected fidelity cut level e g 40 and look on the list of diagnostic species for each of two columns Double click on Gr_3 and the number of columns will increase to 3 Sort the entire synoptic table once more and you can see the diagnostic species for three columns You can repeat this many times and test which number of clusters yields a sufficient number of diagnostic species Pressing the Tree button displays a dendrogram of hierarchical clustering The x axis of the chart displays relev groups proportionally according to the group size and the y axis indicates only the cluster hierarchy 72 Statistics Phi coett A Total time 3 days 12 h 17 min 30 sec Cluster Analysis Dendrogram PC ORD Fig 66 Cluster analysis dendrogram 2 2 1 4 Manual Table Analysis and Import of Results PC ORD cluster analysis from within JUICE i
134. r presentation of spatial data on a national or regional scale In comparison with sophisticated GIS software it gives faster and generally better results because small clear maps with reduced details for publication are usually required for publication JUICE supports easy communication with D MAP The header data must contain LATITUDE and LONGITUDE fields represented as six or seven characters corresponding to degrees minutes and seconds as explained in Section 0 The D MAP software must be installed on the computer The file group setup of D MAP will help to create group JUICE lt title gt with manually defined boundaries and parameters Distribution data will be stored in the JUICE DIS file Parameters such as frame grid spacing and symbol size must be stored in the PAR file in D MAP From JUICE s Options window select the External Program Paths tab and define the D MAP path This will allow you to run D MAP automatically after export selection JUICE can export spatial data for three different parameters for selected species to see the distribution of selected species a group of selected relev s to see the spatial distribution of a vegetation type or all species in the table 44 DMAP JUICE DER File Edit View Options Tools Help BRAY mj oididicl AJAA w Trara 2 S Crambe tataria Records read 13 Records mapped 13 48 20195 N 18 76790 E 48 127 0 N 18 464 4
135. rchy All separators in different levels will be changed into separators of the top level 1 5 5 Gathering Species or Relev s into Blocks Before separators can be used to define species or relev groups it is necessary to gather similar items to the same part of the table While it is possible to manually drag each individual row or column to the appropriate place it is quicker and easier to use colour coding This is a two step process The example below illustrates gathering relev s into a block The process for species is analogous Give the relev s to be gathered a single colour distinct from the other relev s in the table gt JUICE c juice_fin 1 louky_ass_str_del01 wct File Edit Species Relev s Head Sorting Separators Synoptic Table Indicator Values Analysis Table Simulation Help Statistics Total time 2 days 6 h 54 min 43 sec Running number 111 11 1 766 8819555 5552 5749978 77 733311070189777788 208333333377777 Releves 1210 688 5964999488053864995229919222947374224499441674111121112222 Species 606 765434639210087139933236134965567207283206576237864123415690123 Achillea aspleniifol S Achillea millefo Achillea ptarmica Acinos arvensis S Aconitum napelus Adoxa moschatellina Aegopodium podagrari S Agrimonia eupato Agropyron caninum S Agrostis canina Agrostis capillaris S Agrostis stoloni Ajuga genevensis Ajuga reptans S Alchemilla hybri S Alchemilla vulga Alisma gramineum Ali
136. real data From a synoptic table with fidelities see Section 1 11 open the Fidelity measures tab of the Options window Changing the parameters will change the fidelity values in the synoptic table and you will have a clear overview of what happens to your data after standardization Checking the Fisher s exact test check box excludes phi coefficient fidelity values which are not statistically significant they are assigned a fidelity of 0 The level of significance can be chosen 1 10 4 Fidelity Tests This function is included for users who would like to test fidelity measures with theoretical values From the Help menu select Fidelity Measure Test This opens the following window 52 Test Fidelity Measure Yourself Phi coeff A S N no of relev s in the dataset 1000 LI Np no of relev s in the vegetation unit n no of species occurrerices in the dataset r LI LI np no of species occurrences in the vegetation unit Fidelity value 30 589 The value will be copied into the clipboard Fig 45 Window for testing fidelity measures The fidelity measures in JUICE depend on four parameters N number of relev s in the data set N number of relev s in the target vegetation unit n number of occurrences of the species in the data set np number of occurrences of the species in the target vegetation unit The program uses the fidelit
137. relev s by correlating vegetation and site factors and by investigating vegetation patterns Species or relev data can be analysed to find groups Classification or establish a meaningful order 173 Ordination The user can investigate the floristic and ecological composition of relev s as well as occurrence of species The program is distributed as freeware The professional full featured version of MULVA 5 1 is available at http www wsl ch land products mulva 20 2 2 2 1 Initiating Connection MULVA can be connected to JUICE by a process similar to that outlined in Section 2 2 1 1 for PC ORD install the MULVA package and in the External Program Paths tab of JUICE s Options window specify the path to the M51 EXE file 2 2 2 3 MULVA Cluster Analysis from within JUICE As with PC ORD MULVA can be called from within JUICE From the Analysis menu the user selects Cluster Analysis and MULVA The window for parameter settings will appear Cluster Analysis via MULYA Cluster Analysis Setup Data Exported Presence Absence 1 Ordinal Scale 123456789 Percentage Scale Scalar transformation Vector transformation Wo Transformation Wo Transformation Square Ront of Absolute Value Relev Vectors to Unit Length Absolute Value of LOG X Y Except 0 Attributes Vectors to Unit Length LIS Ri Transformation for Correspondence Analysis Adjunt Relev s to 100 Cover Clus
138. rical form Westhoff amp Whittaker 1980 There are five constancy categories widely used in phytosociological literature I 0 20 96 II 20 96 40 IH 40 96 60 IV 60 80 V 80 100 Such synoptic tables are clearer and all columns are easily comparable However differences between categories are hard to judge The differences between two consecutive categories could be anywhere from 1 to 39 Moreover species in the same category could have percentages differing by as much as 19 so a given species may have constancy more similar to species in a different category compared to species in its own category File Edit Species Sorting cuas 2 JUICE c documents and settings lubomir tichy dokumenty O_lubos juice kurz 10 3 2005Ydyje valley1 wct Synoptic Table Table Simulation Help Separator D 5 Categorial synoptic table Number of releves releves 202 Species 631 Poa trivialis Deschampsia cespitos Equisetum arvense Aegopodium podagrari Urtica dioica Geum urbanum Stellaria nemorum Galium aparine Pulmonaria obscurato Sambucus nigra Galeobdolon montanum Phalaris arundinacea Oxalis acetosella Carex sylvatica Glechoma hederaceath Carex remota Dryopteris filix mas Carpinus betulus Hepatica nobilis Dactylis polygama Quercus petraea agg 11 Poa trivialis Frequency 13 Rel o Turboveg No Column 74 Start fli MALACKY By number rounded Gi obsah lto4 doc sx
139. ristic Frequency Index FQI Similarity indices which compare two different relev s without weighing diagnostic species such as Jaccard s Sorensen s Czekanowski Euclidean distance and other qualitative similarity indices give similar results The modified measure of compositional satisfaction proposed by Hill 1989 was selected as a representative index of this type of similarity Unlike the original measure the modified FQ considers percentage frequencies of species occurrences rather than constancy classes This frequency index is defined as FOI 100 Zro Zro Eq 15 ieRAC ieC where FQ is the frequency constancy of species i in a vegetation unit constancy column of a synoptic table Species present in the relev sample plot are indicated as ie A and species present in the constancy column as ieC In the numerator frequencies are summed over all species of the constancy column that are also present in the relev considered while in the denominator the sum is calculated over all species of the constancy column The Frequency Index is defined in an interval 0 100 The Frequency Index seems to be a simple and useful measure of similarity but tests on simulated tables show some situations where it produces results which are in variance with intuitive expectations This index cannot distinguish between relev s composed of only species diagnostic for the vegetation unit and relev s with only widely distributed hence non
140. rth selection will move the display back to the position with the selected species in the tree layer When JUICE is in Header Data Display see Section 1 5 10 the text box can be used to search for relev s that match the text In this case a match occurs if the text appears anywhere in the relev s header data The list will scroll so that the next relev that matches the indicated text is highlighted at the top of the list window Note It is also possible to match the text with any part of a species name The mode of searching can be changed in the Find Species Relev s function section of the General tab in the Options window Options Fidelity measures Synoptic tables Display parameters Separators External program paths Check List Import Dimensions of the program The program is predefined for the maximum size of 30000 relev s and 5000 species This option allows to change the table dimension from 30000 to 65000 relev s and from 5000 to 15000 species Relev s 50000 Species 8000 Confirmation Find species relev s function The function searches a species name or a short cut number or string in short headers Any string part of the header data will be searched in case header data will be displayed Compare string with all species name C Compare string with any part of species name Fig 20 Options General 1 5 12 The Undo Function Some operations colours relev species posit
141. rum sibiricum 9 70 Thalictrum foetidum 1 65 oorooo0oco0coo0oc0c0c 090 Fig 69 The window for interspecific associations Note Species in either list can be selected and assigned a colour which will appear in the table The user can Shift click or Ctrl click to select multiple species The columns in the lists are fidelity measure species name layer species frequency in the data set and frequency of joint occurrence of current and selected species in the data set Note To see the difference between co occurring species and interspecific association consider the following example with Species A as the target species The symbol indicates presence in a relev and a indicates absence All occurrences of Species A co occur with Species B and C giving a co occurrence of 100 with these species Species A also has 100 association with Species C However because Species B has occurrences outside the relev s containing Species A the association between A and B is less than 100 Note also that co occurrence is not a symmetric relationship if B is the selected species A and C have co occurrences with B of less than 50 Species A 4 Species B d4 tdBERBES HT 4 B B B X M B B R RB AGAYMAVYlelll l l l Species C 4 4 2 3 3 Initial Selection of Species into Groups To create an initial group for COCKTAIL analysis select an initial species in the table or assign a colour to several spec
142. rve models I and III the optimum is considered to be the midpoint of this flat segment In the case of model I the constant valued curve the optimum is simply the midpoint of the available gradient values Tolerance is determined similarly to the method used in Schr der et al 2005 It is that part of the gradient where the predicted probability of species occurrence is greater than half of the maximum predicted probability 93 Sphagnum contortum 0 54 g optimum 04 5 90 S 0 3 4 pi2 o 74 1 74 0 24 min 4 max Ks 5 48 f 6 31 j int li 4 s j 0 83 0 0 4 hor TY T T T T T 3 4 5 6 7 8 water pH WC 56 1 5 8 18 7 10 9 4 BG 46 8 18 1 74 9 6 18 1 Fig 81 Optimum and Tolerance calculated by the species response curve algorithm Hajek M 2 8 2 Installation of the Function in JUICE Calculation and graphing of response curves is done in the R software environment R Development Core Team 2005 but it is run under the complete supervision of JUICE The R package must be installed together with the required R libraries and the text file containing R script JUICE must be updated if the version is older than 6 3 108 The R software package is free open source software available at http cran r project org bin windows base HOF models in R are calculated using the gravy library written by Jari Oksanen This library is not included in standard version
143. s Phi cott A Cluster Analysis Dendrogram lt TUI alll LT 1357911 1416 1819 22 24 25 2832 363942 4445 46 a a 2a 56 e paa 60 uH 717376 78 8084 88 9092 i 100 2468 12 1517 20 2630343740 43 6569 7275 798286 89 9194 r Frequency Rel o 1 Alnus glutinosa 4 Turboveg No Fig 27 Dendrogram of PC ORD cluster analysis 1 6 4 Autorepeat Function When a new table is imported into JUICE it is often necessary to merge species to aid analysis See Section 1 6 1 The user may also wish to delete species that are not of interest to the current project See Section 1 6 2 When new relev s are added to the original data set e g in TURBOVEG it will be necessary to perform the same mergings and deletions in the newly exported file The Autorepeat function available from the Species menu can be used to avoid the tedium of repeating the same editing operations in multiple files When the Autorepeat Function window opens press the Add functions from WCT file button This will allow you to select a previously edited file as a model for how the current file should be edited Autorepeat Function Current step CONCAT Achillea millefolium agg CONCAT Achillea millefolium agg Achillea setacea Agrostis canina
144. s a source for calculation of crispness of classification 4 Call the Crispness Of Classification function from the Synoptic Table menu This window which is also always on top of the main window will appear 66 x AR foec A x col K Statistics Phi cost A a Total time 1 day 21 h 14min 40 sec Percentage synoptic table Humber of releves releves 471 Species 465 Eriophorum vaginatum Sphagnum rubellum Limited list for species of colour List Of Values Actual Average Crispness Value 34 098 Cirsium rivulare Eupatorium cannabinu 107 Rel o Turboveg No Column NN 9 i FUR gt wee o o Brux LL T BR acdsee fo iB Interne fc Microso Q CorelPHOT Sy 2 Visual B C3 uce lona CS 26x prae Qum 20 22 11 S Agrostis canina 0 Fig 61 The sorting function and the Crispness Of Classification window overlaid on top of a synoptic table 5 Press the Refresh button The program will calculate the appropriate value for the selected number of clusters 6 In the cluster analysis window select a new number of clusters The synoptic table will change Press the Refresh button in the Crispness Of Classification window to calculate a new crispness value All values are saved onto the clipboard 7 Iterate these steps with increasing numbers of clusters Note 1 The progr
145. s a spreadsheet table Whereas the MS EXCEL CDY export has header information at the end this export has header information at the beginning Furthermore the former format presents header data by listing one relev s data per row whereas this format lists one field per row with each row containing each relev s data value for that field 42 Table Export Species below Exported relev s the table e Rare species will be moved below the table into the text form Please define their maximum frequency Red Hone a Inless than 2 relev s f Blue In less than 3 relev s C Sea green In less than 4 relev s Green Violet Grey All Export separators File format Species exported MS WORD SDF em MS EXCEL CD with non zero value NM C a Cancel s PACTES usetul e g tor large tables which i TEXT FILE need to be printed in sections and CSV HEAD TABLE completed in paper form Continue gt gt gt Fig 37 Window for table exports It is possible to restrict export to relev s of a certain colour The Species Exported box gives the option of exporting all species in the table or only those species which appear in the selected relev s Note Exporting all species can be useful To export a large table for printing select a screen full of relev s at a time and export all species When the resulting files are printed out the species lists will be complete and the printouts can b
146. s group were 5096 of the total data set size and the other three groups together were also 5096 We can then compute a fidelity measure for the second vegetation unit as though it were 50 of the total data set size and so on Such standardizations make sense in combination with the phi coefficient but not so much with ump or other statistical fidelity measures because data standardization violates their relationships to statistical significance A scheme of different standardizations of the size of vegetation units is presented in the next figure Each line represents a data set segments are vegetation units labelled A B C and D and segment length corresponds to the number of relev s in each group The thick part of each segment represents relev s with occurrence of the given species the thin part represents relev s where the species is absent In figures b to d the four lines represent the four standardizations used for calculation of species fidelity to the target vegetation units A B C and D 49 a original data 8 ee 8 quu 8 5 05 QANENRUNER b standardization B A C D of the size of the A ESS GL target site group jamma C A B D only D A B m c standardization 8 1 8 mam 88 im of the size of all C C site groups DTT a a jm target group being ee 7 uu essent of different size D A B than the others jpe a p E
147. s limited by the settings available from the input window Fig 64 Manual data set classification is less convenient but it can be used in cases when it is not possible to connect PC ORD and JUICE or when the user wants to use a PC ORD function not supported by JUICE To classify data manually do the following note that steps 1 and 10 are in JUICE while steps 2 through 9 are in the PC ORD program 1 From JUICE export the whole data set or a selected part to a CC file Run PC ORD Import the CC file through the TWINSPAN DECORANA filter Apply data transformation if necessary Select menu Groups and Cluster analysis and choose the type of analysis In the Cluster Setup window check the Add group membership variable to Second Matrix box 7 In the Group membership variable window enter the Group level and check the Write all higher level groupings box 8 Create the Second Matrix 9 Export the Second Matrix as Comma Separated Values Spreadsheet 10 Back in JUICE from the Sorting menu select Sort relev s by clusters and PC ORD Then proceed as in Section 2 2 1 3 S vro Es 2 2 2 Cluster Analysis via MULVA MULVA 5 is a program package designed to apply multivariate statistical methods to vegetation and site data as a means of investigation in plant ecology A complete description of MULVA S is available in Wildi amp Orl ci 1996 MULVA 5 enables analysis of
148. s manual 1 8 2 9 Importing External Short Header Data External header data can be imported from a text file From the File menu select Import and Short Headers The data file must have a column of relev numbers 6 characters paired with a column of short header values 6 characters Short header files in this format can be exported from JUICE from the File menu select Export and Short Headers See also Section 1 9 9 The Import Short Headers window is shown below Import Short Headers C Documents and Settings Lubomir Tich D okumenty 0_lubos Prace WUICE Clanky JUICE manual B 3Xsh expo txt i for relev s B wis j Source positions Destination positions Fig 34 Window for import of short headers The Open button opens a browse window for finding the file with appropriate header data The colour in the for relev s field restricts which short headers will be updated The Source positions fields are used to specify the beginning and end of the header data in the source file while the Destination positions specify the position in the short header In the example above 2 digits the fifth and sixth characters after the relev identification number will be stored at the first and second position of the short head The third through sixth characters of the short head will not be overwritten Note If you respect the convention
149. s stone setas stone setas sense tnu 10 1 5 1 Table Window Comrpohnents eed eene ie ede ede Eee nie dehet de deeds aede 10 1 5 2 Mouse Keyboard Functions iss eee ies eie edid eee ne ertet de ee REA 11 1 5 2 1 Functions sorted by displayed objects sese nnns 11 1 5 2 2 Mouse Functions Sorted by Similar Functionality sse 15 1 5 3 COO UT S eter or teas ae Yate ae dos ele RERO HE ERR Fee DOES Oe eh I e ea 16 1 5 4 SEP LMOTS nE e EE 16 1 5 5 Gathering Species or Relev s into Blocks sss 16 1 5 6 Relev Overview Displays eee tegere et soe ee vies ae ees se esque es 18 1 5 7 Editing Species and Header Data E EE C AINE a dente 18 1 5 8 Screen Options ooo ee reU REST SU GR IS OS TRU SOR HE AU I T INE TE ERNBETS 19 1 5 9 Defining Scales or sett tes ese a de e RR ERUNT DIS 20 1 5 10 Header D ta 5 inert oet petere ed tsuneo a De eee eo exe steel ie eels Exe Toe rne vede PE pen 21 1 5 10 1 Selecting Relev s by Header Dat eee eS ee ts Redes hee E ees 22 1 5 11 Sr dun s 23 1 5 12 The Undo Functions oenen dt oer eee A Ita De ere ERNER HE Re eate Te koe E E 23 1 6 Editing Lables e 24 1 6 1 Merging Species e eR Osee ie a e vu Ate a TB RE RU NEUES 24 1 6 2 Deleting and Undeleting Species and Relev s sss 25 1 6 3 Sorting Species Species Data and Relev s sese eene 25 1 6 3 1 Sorting Sp
150. saryk University Brno Czech Republic Free distribution of the program via internet has been available since 2001 This manual describes the possibilities of the program version 6 3 66 Newer versions may differ 1 2 Copyright Information JUICE is freeware which can be freely distributed as an original package The program download on www sci muni cz botany juice htm is without registration No official warranty or support is provided Questions not covered by this manual or the FAQ can be sent to tichy sci muni cz In publications or reports containing outputs from the program the paper with the basic information on the program Tichy 2002 should be cited 1 3 Installation This chapter explains how to install and initialise JUICE 1 3 1 Content of the Installation Package The 4MB installation package can be downloaded from the internet address http www sci muni cz botany juice jc05_ins htm There are two hyperlinks a full installation download and a JUICE EXE file download New users must use the full installation for correct installation of all program components The installation package contains these files JUICE EXE the program file This file can be replaced directly without new installation by a newer version of the EXE file ELENB TXT Ellenberg indicator table Ellenberg et al 1992 formatted as simple text The file covers six main environmental factors light temperature continentality humidity pH Ca and nitr
151. sence absence versus Average cover from the Fidelity measures tab of the Options window Option A is to use the given data without standardization Option B standardizes the size of the vegetation unit relative to the rest of the dataset Option C has two steps 1 it sets the target group to the indicated size and then 2 sets all the other groups or all but the last to be the same size as each other The picture on the right side of the frame graphically represents the chosen standardization applied to a dataset divided into four vegetation units Warning Because the Dufr ne Legendre Indicator Value implicitly standardizes the data set it is only available with option C A side effect is that when the user scrolls through the list of measures JUICE automatically switches to option C when Dufrene and Legendre is highlighted If a different standardization is desired for a different measure it must be re selected manually Similarly highlighting Dufrene and Legendre will cause JUICE to switch from Average cover to Presence Absence data Note 1 The data standardizations are available only for synoptic tables and related functions thus they are only applicable to classified or partly classified datasets All other functions using fidelity as a statistical measure such as COCKTAIL classification and Interspecific associations use the original data Note 2 You can test the standardization with
152. sma lanceolatum Alisma plantago aqua Allium angulosum Allium carinatum Allium oleraceum Allium scorodoprasum Allium vineale Frequency Rel o 3 Row Turboveg No 403026 Column we 1 Achillea aspleniifolia Fig 10 The table with selected relev s Gather the relev s using the arrow buttons on the Icon Bar or the Move Relev s function in the Sorting menu Specify whether the relev s are to be gathered to the left or to the right The illustration below shows relev s that have been gathered to the left ACE peN juice fim lowky ans str det wet Running number 11 11 1 i 1 766 88 555 4 107018977 337195552579978 77733317788208333337777 Releves 1210 688 5948803414737422491112649995869952299922294944674112111222 Species 606 765434608719360728320677340392103932361349556725623864121569123 Achillea aspleniifol S Achillea millefo leelesloreee 2 811692182 92121411812242 224218 192 18111812 24 Achillea ptarmica i Acinos arvensis 1 S Aconitum napelus Adoxa moschatellina Aegopodium podagrari 5 Agrimonia eupato Agropyron caninum 5 Agrostis canina Agrostis capillaris 2 1 2 2 2231 142 221 312222 5 Agrostis stoloni Ajuga genevensis Ajuga reptans eer 2 1 1 2 5 Alchemilla hybri 2 1 S Alchemilla vulga tee92 12 2223 2121 11221121 23323 1 Alisma gramineum Alisma lanceolatum Alisma plantago aqua Allium angulosum Allium carinatum Allium olerace
153. species groups correspond to the concept of sociological species groups Doing 1969 and often they are closely related to the groups of diagnostic species for particular vegetation units as recognized in phytosociological literature Extraction of each group starts with one or a few species pre selected by the researcher Other species with the most similar distribution across the relev s of the database are added stepwise to this starting species or species group In JUICE co occurrence tendency of species is measured by the phi coefficient of association Sokal amp Rohlf 1995 Chytry et al 2002 Unlike Bruelheide 1995 2000 who used a fully automated process of species group optimization the approach in JUICE allows for more manual control with the aim of arriving at ecologically more coherent species groups After selecting a starting group of two or three species the user may calculate a fidelity coefficient of association between each species in the data set and the group of relev s that contains the starting species group Of the species not belonging to the species group the user will usually choose the one with the highest phi coefficient and will include it in the group as its next member Sometimes it may be intuitively more correct to include the species with the second or third highest phi coefficient particularly if the species with higher fidelity coefficients have already been included in another species group or have several tim
154. starting path with INI file and TWINSPAN must stay open for program outputs 1 3 2 Computer Configuration The program is written for the WINDOWS operating system in English US format The pre defined decimal delimiter must be a period The program will automatically try to convert a decimal comma to a decimal point when the program starts A previous selection of the decimal point is restored after the program termination Some known problems occur with Asian formats of the operating system Therefore if you encounter problems such as absence of communication with TWINSPAN or unreadable export files try setting Regional Settings to English US The program has no special demands on computer hardware However listing through large tables may be slower on older computers 1 3 3 Program Settings and INI File JUICE saves the user s settings and restores them the next time the program is used The settings are saved in the JUICE INI file which can be found in the same directory as the program This file is automatically generated the first time the program is run and it is updated during table operations If the file is deleted from the JUICE directory the program will use predefined settings Note Each line of the INI file contains a parameter name and a parameter value separated by a double colon thus it is possible to edit the file manually However almost all values can be managed directly from the program Manually al
155. t vegetation unit Note The Dufr ne Legendre Indicator Value s relationship to independence can be seen by comparing Fig 40 a or b with Fig 41 In each figure the diagonal from the lower left to the upper right is the case of independence the species is as frequent inside the vegetation unit as it is outside the vegetation unit Note that this diagonal corresponds to the 0 isoline in Fig 40 but it is crossed by the isolines in Fig 41 IndVal assigns this diagonal a value of p 2 where p n N as can be seen from where the 0 2 and 0 4 isolines cross it 48 d Fisher s exact test calculates P f o 2 nj the true probability of obtaining the observed number of occurrences of the species in the vegetation unit under the null hypothesis of independence It may yield very small probability values including those smaller than 10 99 which are difficult to cope with in practical work For this reason logio P f o np is a more practical quantity to use for a measure Fisher s exact test is used either as a stand alone fidelity measure or as a correction for calculation of missing information on statistical significance if the fidelity synoptic table with phi coefficient is displayed Note Negative values of fidelity represent information which is not important in the context of most studies Negative values are displayed only in the Interspecific Associations window More on interspecific associations will be included in la
156. te ornamental Basella rubra casual l deliberate food Spergula arvensis naturalised pi 5 accidental Species name Quercus rubra Mark gt species with the value 1 with the colour Cancel all P Do not import data ances ma zaa Fig 29 Window for import of species data from external file Species names in the external data file and the current table must have the same nomenclature The Parameters for species selection and Species information bounds must be entered manually The first number tells JUICE where to find the first character of the field and the second number tells JUICE where to find the last character The scale displayed above the file excerpt can be used as a guide For example the column below the first 1 corresponds to character number 10 and the column below the first 2 corresponds to character number 20 Warning 1 Make sure that the range entered under Parameters for species selection is large enough to include the longer species names Ideally the Last character of the species name should be the character immediately preceding the first column of data If JUICE seems not to be loading in data for species with longer names this indicates the entered range may be too small Of course specifying too large a range will also cause problems Warning 2 To ensure that the external data file will be readable separate columns with spaces and not with tabs Your text editor shoul
157. ted Suppose for example that the model file had Quercus robur in two layers that were subsequently merged but the current file has Quercus robur in three layers After performing the merge step from the model file only two of the instances of Quercus robur will be merged The third must be merged manually The optimal application of this function is to perform the same merging deleting or undeleting criteria in different sub sets of one large data set or in the same data set after slight modifications of the source database 1 7 Species Data The second column with the light grey background can hold additional species data such as layers ecological characteristics or biological information which can be used in sorting and analysis Section 1 6 3 2 explains how to sort species by species data Species data can have up to 50 characters however not all the characters are displayed By default the Species Data Column has a width of three characters It can be enlarged in the Display Parameters tab of the Options window as described in Section 1 5 8 JUICE can write some information to the Species Data Column automatically From the Species menu select Species Data The functions available are explained below 1 7 1 Layers A species s layer is expressed as a number from 0 to 9 It is recommended that you use the same convention as the TURBOVEG database program 31 0 not defined 1 tree layer high
158. ter versions of this manual They are hidden in synoptic tables 1 10 2 Fidelity Measurements for Vegetation Units of Unequal Size All the fidelity measures included in JUICE except the Dufr ne Legendre Indicator Value are affected by the relative size of vegetation units Unfortunately nearly every classification of relev s yields vegetation units of different sizes To remove the dependence of the fidelity measures on the vegetation unit size it is better to make a virtual standardization of the target vegetation unit V to a new value hereafter called N which is constant for all vegetation units within the data set without changing species frequencies within and outside each target vegetation unit i e the quantities n N and n np N N respectively For example in a data set with two vegetation units we can set N N 2 i e equal to half of the total size of the data set which enables us to compare the two vegetation units as if they were of the same size Thus the resulting fidelity values are less dependent on the sampling effort and data set structure Similarly in a data set with four vegetation units we can set N N 4 However it is not necessary that the sum of N values for all vegetation units within the data set equal N the total size of the data set For example in a data set of four vegetation units we can set JV N 2 and compute a fidelity measure for the first vegetation unit as if thi
159. tering the INI file may cause problems If you experience difficulty with the INI file simply delete it JUICE will generate a new one in the correct format 1 4 Data Import The first step in working with JUICE is to open a table JUICE does not support direct storage of phytosociological relev s Therefore all relev s must be entered using other software such as TURBOVEG and exported as a table that can be read by JUICE JUICE accepts several different file formats 1 4 1 XML Format This format is useful for importing source data from the TURBOVEG database program Hennekens amp Schamin e 2001 The XML file contains full information about species names synonyms cover and header data and it is not necessary to create any other file with additional information The file structure is rather complicated and manually entering data in this file format without TURBOVEG is not recommended TURBOVEG XML Import File Step 2 2 XML format file C Tmpitest xml Header data Relev number i Relev number County code Biblioreference No table in publ Add all fields No relev in table Cover abundance scale Project code Date year month day Syntaxon code Remove all fields Relev area m2 Finish Fig 1 XML file import The XML file contains all header data but JUICE allows reduced header data information The user must select the fields to be imported from the left list box using the Add
160. ters Select number of clusters MI Interval 2 105 Ressemblance Function Group Linkage Method Cross Product Cross Product Centered Gower Index Covariance Euclidean Distance 7 Correlation Coefficient Chord Distance Complete Linkage Ochiai Coeficient Manhattan Distance Min Variance Clustering Van der Maarels Coef FSPA Distance Ward s Method Single Linkage Path of the output file CAMULVA nulya dat Cancel Continue gt gt gt Fig 67 Parameter settings for cluster analysis with MULVA Cluster analysis can be applied to the entire data set or to a selected part Cover data may be exported as presence absence ordinal or percentage values In the case of semi quantitative or quantitative data scalar and vector transformations are available MULVA cluster analysis calculates only a selected number of clusters without any information about hierarchy JUICE is able to eliminate this disadvantage by repeatedly calling the procedure with an increasing number of clusters Check the Interval box This produces a dendrogram similar to the one discussed in Section 2 2 1 3 The MULVA parameters window provides several resemblance functions distance measures and group linkage methods 2 2 2 3 Results and Cluster Tree The tools for data set classification described in Section 2 2 1 3 can also be used with MULVA 74 2 2 2 4 Manual Table Analysis and Import of Results It is also possible to an
161. time of calculation After pressing the Continue button it is not possible to cancel the process Warning 2 The average fidelity value is written into the species data field This enables the user to see the value but it overwrites any existing species data JUICE gives no warning before it does this 1 6 3 3 Other Relev Sorting Functions Sort Short Headers is useful when important relev data are stored in the short headers Short headers can contain ecological information about the relev relev number number of selected species or other information To write information to the short headers select Store Values To Short Headers from the Head menu This information is described in detail in Section 1 8 2 28 Note Short headers are limited to 6 characters Numbers should be sorted in numerical order 1 2 3 11 12 13 21 22 23 while text strings should be sorted in alphabetical order If the example sequence above is sorted in alphabetical order the result is 1 11 12 13 2 21 22 23 3 Sort Relev s by Header Data allows sorting according to any field in the header data This function includes the option to write the initial characters of the selected header data field to the short headers so the user can see the values according to which the relev s have been sorted and the option to add separators after each group of relev s with identical values Sort Relev s By Selected Header Fie
162. tion Folia Geobot Phytotax 32 41 46 Bruelheide H 2000 A new measure of fidelity and its application to defining species groups J Veg Sci 11 167 178 Chytry M amp Tichy L 2003 Diagnostic constant and dominant species of vegetation classes and alliances of the Czech Republic a statistical revision Folia Fac Sci Nat Univ Masaryk Brun Biol 108 1 231 Chytry M Tichy L Holt J amp Botta Dukat Z 2002 Determination of diagnostic species with statistical fidelity measures J Veg Sci 13 79 90 Czekanowski J 1913 Zarys metod statystycznych Die Grundzuge der statischen Metoden Warsaw Doing H 1969 Sociological species groups Acta Bot Neerl 18 398 400 Dufr ne M amp Legendre P 1997 Species assemblages and indicator species the need for a flexible asymmetrical approach Ecol Monogr 67 345 366 Ellenberg H Weber H E D ll R Wirth W Werner W amp Paulissen D 1992 Zeigerwerte von Pflanzen in Mitteleuropa Ed 2 Scripta Geobot 18 1 258 Hennekens S amp Schamin e J H J 2001 TURBOVEG a comprehensive data base management system for vegetation data J Veg Sci 12 589 591 Hill M O 1989 Computerised matching of releves and association tables with an application to the british national vegetation classification Vegetatio 83 1 2 187 194 Hill M O 1973 Diversity and evenness a unifying notation and its consequences Ecology 54 427 432 Hill M O
163. to select a model and test it on sample data The user may then choose whether to write radiation or heat load to the short headers For relev s which lack information about latitude slope and aspect a null value is written Note For more information read the original paper of McCune and Keon 2002 Latitude should be included into header data See the Section 0 1 9 2 7 Sum Average Minimum Maximum and Multiplication of Species Data There are several functions for summarizing numerical species data in the short headers Average of colour Species Data calculates the average over all species of the indicated colour in the relev Sum Maximum and Minimum functions are similar Multiplication Of Species Data Values is an average calculated as follows 38 SD SD SD E mean SD Eat where n is the number of species in the relev SD is the species data value for species and mean SD is the average of species data values taken over all species in the table These functions apply to the species data currently in the Species Data Column See Section 1 7 for more information about species data including how to write values to the Species Data Column These functions are only useful for analysing numerical species data 1 8 2 Ellenberg Indicator Values Ellenberg indicator values can be written to the short headers More information on Ellenberg indicator values will be included in later versions of thi
164. ty measure see Section 1 10 1 is calculated for each pair of species which gives information on their reciprocal affinity in the data set First select the species to be analysed by clicking on it in the table Then from the Analysis menu choose Interspecific Associations This opens the Interspecific Associations window shown in Fig 69 The window shows a list of species positively associated with the selected species and a list of those negatively associated The list of positive associations can be exported to the current RTF export file 77 Interspecific Associations Circaea alpina Positive association Number of relev s 221 species Total dataset Maximum value Diplazium sibiricum Oxalis acetosella Athyrium filix femina g x Phegopteris connectilis Mark species in the table Plagiochila porelloides Dryopteris expansa Paris quadrifolia Thuidium philibertii Add to RTF file S Ribes nigrum Gymnocarpium dryopteris Rhytidiadelphus triquetru 8 Abies sibirica black Only selected species Cinna latifolia Negative association 806 species Mark species in the table E S Caragana pyguaea e E Lare LEE Q i oa sect Stenopoa 6 1285 Artemisia frigida 6 103 Galium verum 6 102 Dianthus versicolor 6 96 Veronica incana 9 91 Potentilla acaulis 6 90 Orostachys spinosa 6 83 Aster alpinus 3 81 Koeleria cristata 4 80 Kitagawia baicalensis 6 98 Achnathe
165. ty measures Synoptic tables Display parameter si Text parameters The table design 20 Text size 135 Zea mays Text lenght 20 Bold style of the text iv Background contrast Actual scale Braun Blanquet Old Scale Normal Select one from usual scales for the display of your data Non frequent scales are also accepted Lenght of additional species information The field Species data represents species dependent information about each species which can be modified imported exported and analysed in the program e g layer ecological information biological data The program will display 3 characters Fig 15 Options window Display Parameters Note The width of the Species and Species Data Columns can also be modified using the markers at the top of the table Text size can also be modified using the two text size icons on the Icon Bar 1 5 9 Defining Scales All cover data are saved in the form of percentage numbers The program accepts integer numbers from 0 to 100 Each number from 1 to 94 can be assigned a character while numbers from 95 to 100 are assigned the same character The program has four predefined scales Braun Blanquet Old Braun Blanquet Old and New Ordinal and Presence Absence All other scales must be defined by the user The scale is selected in a Braun Blanquet Old Scale Hormal Braun Blanquet Old and Hew Scale Hormal Braun Blanquet Old Scale A
166. ulate only uniqueness for all columns Calculate all columns Values are copied into the clipboard Fig 59 Uniqueness and Asymmetric Similarity Indices calculation The list of species compared is limited to those with fidelities higher than the defined threshold value This value depends on the selected fidelity measure and method of data standardization The vegetation unit can be selected with the scroll bar After pressing the Calculate one column button the calculated Uniqueness Index and a list of Asymmetric Similarity Indices are displayed This information is also saved onto the clipboard Note The Calculate only uniqueness for all columns and Calculate all columns buttons are for copying the indicated information to the clipboard from which it can be pasted into another program The information actually displayed in the window will be the full information for the highest numbered relev group 1 11 14 Average Values of Constancy Columns The Average Values Of Constancy Columns function of the Synoptic Tables menu calculates the average for each synoptic table column The displayed values relative frequency fidelity average cover etc are averaged and these averages are displayed in a separate window Average Of Positive Fidelity Values t Fig 60 Average values of constancy columns The results can be saved onto the clipboard for further use 65 1 11 15 Crispness of Class
167. ull species names The program uses a species list file in simple text format Such a file can easily be exported from TURBOVEG From the Manage menu in TURBOVEG select Species Lists and Edit Enter the name of the species list and select Export and Limited List For JUICE It can also be created manually as a comma delimited file or a file with a fixed length for each line Examples are shown below Format 1 I I T 1 ABIEALBAbies alba 12251ABIEGRAAbies grandis 2 ABIE SPAbies species 4 ABITASAAbietinella abietina var abietina 5 ABITASHAbietinella abietina var hystricosa 3 ABITABIAbietinella abietina Each column has the same number of characters in each line The first line defines three fields with 5 7 and 50 characters Format 2 1 ABIEALB Abies alba 12251 ABIEGRA Abies grandis 2 ABIE SP Abies species 4 ABITASA Abietinella abietina var abietina 5 ABITASH Abietinella abietina var hystricosa 3 ABITABI Abietinella abietina Data are in three comma delimited columns Format 3 ABIEALB Abies alba ABIEGRA Abies grandis ABIE SP Abies species ABITASA Abietinella abietina var abietina ABITASH Abietinella abietina var hystricosa ABITABI Abietinella abietina Data are in two comma delimited columns Note The species list included in the JUICE installation package is useful only for central European users who are using TURBOVEG wit
168. um Allium scorodoprasum Allium vineale Fig 11 The table after moving all selected relev s to the left Note that JUICE does not gather all the relev s to the leftmost edge of the table They are simply gathered to the leftmost or rightmost relev of the selected colour To move the block of relev s to the left use this trick 1 Drag the leftmost relev in the block to the place the block should be 2 Gather the relev s to the left again This will move all the relev s to the left To move them to the right move the rightmost relev to the desired position and re gather to the right 18 If the program seems to not respond to an attempt to gather relev s or species make sure the correct colour is selected in the Icon Bar JUICE only gathers relev s or species of the selected colour JUICE can also group species and relev s automatically without using colours The Sorting menu contains several other options for sorting species and relev s See Section 1 6 3 1 5 6 Each relev in the table can be displayed in condensed form by double clicking on it See Section 1 5 2 Species in the relev can be sorted by layer cover alphabetical order or current position in the table When the relev display is opened or its sorting method is changed the relev is copied into the clipboard memory the user can paste this information into a text editor or other program Relev Overview Display JU
169. vegetation units in the data set If diagnostic species for a few similar community types are determined using a large database it may be useful to perform fidelity calculations with a data set that also includes relev s from other unrelated community types Chytry et al 2002 This approach finds diagnostic species that are of more general validity because they are tested against the background of other community types in the same geographical area The additional relev s are usually treated theoretically as a single vegetation unit Standardization to the size of other vegetation units would greatly and undesirably reduce its effect However its size can be held constant while the size of the other groups is standardized The phi coefficient applied to a data set with vegetation units standardized to equal size is independent of the actual differences in size of individual vegetation units However it depends on the standardized size of the target vegetation unit N which may be either equal to the 50 size of the other vegetation units or set to any arbitrary value between 1 and N 1 Setting the standardized size of the target vegetation unit V5 to a higher value gives a higher weight to common species and their frequency in the target vegetation unit By contrast setting V to a lower value gives a higher weight to rare species and to the differences in species frequency within and outside the target vegetation unit Changing th
170. with values higher than the cut off level are held fixed at the top of the table The remaining species are sorted in the second column 4 The process repeats for each column until the entire table is sorted The sorting function Entire dataset without last column sorts the table similarly except that the last column is not sorted 61 Note The actual sorting algorithm is more complicated than that described above Species that have values above the cut off level in two columns are moved below the species that exceed the cut off level in only one column These are followed by species that have values above the cut off level in three columns four columns and so on The result will be a table with uniquely diagnostic or constant or dominant species in blocks at the top followed by blocks of species which are diagnostic for multiple groups 1 11 12 Analysis of Synoptic Columns Combining with Exporting the Results The program can extract diagnostic constant and dominant species from each synoptic column This function is widely used for analysis of classification results and interpretation of vegetation clusters lAnalysis Of Synoptic Table E Fidelity threshold A e Frequency threshold 1 100 e Cover threshold 1 100 Diagnostic species 5 Constant species 8 Dominant species 100 0 Carex canescens 6 100 Lysimachia vulgaris 6 100 Molinia arundinacea caerulea 6 82 4 Molinia arundinacea caerulea 6 100 Betula
171. y measure that has been selected from the Fidelity measures tab of the Options window See Section 1 10 3 The current measure is displayed in the Test Fidelity Measure Yourself window and on the Option Button on the Icon Bar Note 1 Standardization methods for the phi coefficient available from the Fidelity measures tab do not affect this calculation Note 2 The fidelity value is copied onto the clipboard and it can be pasted into other programs 1 10 5 Tests of Data Structure Using Different Types of Standardization JUICE also has a function for visualising the effect of data standardization on the phi coefficient From the Help menu select Advanced Fidelity Measure Test 53 The influence of data standardisation on fidelity values phi coefficient xi Data standardisation paN Show np Hp nnn for all species at each column Show fidelity cut levels 0 2 np N Np 1 Fig 46 Window for visualising The influence of data standardisation on fidelity values The window contains a contour diagram depicting the dependence of the fidelity measure phi coefficient on the relative frequency of species occurrences within vertical axis and outside horizontal axis the target vegetation unit This diagram depends on the relative size of the vegetation unit a parameter which begins at 10 and can be adjusted by the user The curves are phi value isolines of 0 9 0 8

JUICE. To manually

Contents

Download Pdf Manuals

Related Search

Related Contents