UNICOR USER MANUAL - Computational Ecology Laboratory
Contents
1. 7z exe Pkunzip Unzip or an equivalent Seven Zip 7Z exe is highly recommended since it can handle all common formats on Windows MAC OS X and Linux On Windows it is best to setup a project specific modeling subdirectory to perform your modeling outside of any folder that has spaces in its name like My Documents 2 2 3 Install UNICOR UNICOR user manual 8 Next install the UNICOR software itself by unpacking the zip archive supplied At this point you should be able to execute the supplied test inputs Alternatively and only for Windows operating system you may double click on the unicorn _setup ex xecutable file for an automatic download 2 2 4 Optional Python Extension Modules As UNICOR is supplied in the archive it does not require any additional contributed Python modules to run However several additional Python modules are needed if you want the following functionality NetworkX is required for graph theory metrics and can be obtained from http networkx lanl gov wxPython is required to run the GUI and can be obtained from http www wxpython org 2 3 Example run 2 3 1 Command line run The example run is for 10 points representing individuals on a Euclidean distance resistance surface To run the following example follow these steps 1 Double check the UNICOR source UNICOR additional packages and UNICOR example files are in the s2. SS Wee OS Shera OF Ge Oe are WES Od ESOS 8 2 3 L Command Laine rune jf ei A wo Sate oa 8 PUT Re eee ete Hen eds we AA eee A A TA aha hae Bie Suey Boe 9 2al Resistance Grud wise dh twine OAM eee ad tiene E Ae oe os 9 Sear AICA CUNAS A Bid en oe ard Yara a ASA 9 343 Thiresho LatnG ist 6 5 sae A ie alee A Wee ls ed ad tees 9 34 4 Numbet of procesor S sssr ene ear ee Sale a Sees 9 3 5 Kernel density eStimat LON li ad Se ee ee 9 OU NS eh aero hear a E nits Wages REN 10 General ISSUES edt edie SA e Sees Sle ee as aie y e 1 Dig Al HOw tO Obtain UNICOR Satie diene Rieger tne See eee hb o fa 11 Diz Debugging and troubleshooting sundae as Sethe te bs 11 D3 UNE COR VAMEEIELON Sie Se oad A RE OER AA ES 1 5 4 Howto Gite UNDCOR iSu bad aj gui ei Pas ee OS 575 DAS Glave A ae Ee A INS E 12 Reference scsi A bois at A o a is ae 12 ACKNOWLEDGEMENT Sissi deseo Boas A sod wo eer e Ree BR Blo elas dupe vee eae 14 UNICOR user manual 3 1 Introduction Habitat loss and its effects on populations of vulnerable species is among the most urgent problems in conservation ecology It is critical that managers and scientists have effective tools to evaluate the effects of landuse and climate change on the extent and connectivity of populations To address this need we introduce UNIversal CORridor network simulator UNICOR a species connectivity and corridor identification tool UNICOR applies Dijkstra s shortest path algo
3. of expected movement rates for every pixel in the study area extent rather than only for a few selected linkage zones Compton et al 2007 Second scale dependency of dispersal ability can be directly included to assess how species of different vagilities will be affected by landscape change and fragmentation under a range of scenarios e g Cushman et al 2010a Third it is computationally efficient enabling simulation and mapping across vast geographical extents for a large combination of species e g Cushman et al 2010b Cushman et al 2011 2 Getting started 2 1 Dependencies 2 1 1 Baseline Requirements UNICOR requires the Python2 6 x interpreter NumPy package and SciPy package Several optional Python module packages if enabled facilitate additional UNICOR functionality Remember that Python modules usually require particular Python interpreters so be sure the version ID for any external Python module or package e g NumPy or others matches the version of your Python interpreter normally v2 6 X 2 1 2 Python on Non Windows Platforms Some common computer platforms come with Python installed These include MAC OS X and most Linux distributions To determine which Python a MAC or Linux workstation has installed start a terminal console and enter python You ll see the version number on the top line enter Control D to exit Replacing a
4. through outputs of graph theory metrics e g density number of nodes radius etc and connectivity outputs that can directly input into popular landscape pattern analysis programs e g FRAGSTATS McGarigal et al 2002 Table 1 UNICOR inputs and outputs with description and dependencies Input Name Default Description Dependency Example Input Grid Filename small_te The filename for the resistance surface st rsg in ascii format with header file any file extension is acceptable must be space delimited XY Filename small t The filename for the individuals with st _10pts x y locations any file extension is Xy acceptable must be comma delimited Use ED Threshold False Option for using Euclidean distance thresholding ED Distance 50000 If Use ED threshold is True then the Euclidean distance in map units to apply to the x y point locations Edge Distance 100000 The resistance distance threshold in terms of edge distance to apply to the path lengths Edge Type Normal normal All pairwise cost distance paths with no thresholding threshold apply the threshold value given by Edge Distance to all paths all paths calculate resistance UNICOR user manual 11 kernels for all points T
5. UNICOR user manual 1 UNICOR USER MANUAL Version 2 0 Updated 2012 02 04 Authors E L Landguth B K Hand J M Glassy S A Cushman and R T Carlson 1 University of Montana Division of Biological Sciences Missoula MT 59812 USA 2 Lupine Logic Inc Missoula MT 59802 USA 3 U S Forest Service Rocky Mountain Research Station 2500 S Pine Knoll Dr Flagstaff AZ 86001 USA Table of Contents 2 od 1 How does UNICO UNICOR user manual 2 Introduction A bala E A AA AA E E E O A E 3 Tat What can UNICO R WOT Ka ea ale aaa en a 3 1 34 Whats new Tn Version AD oi Bale Ee lee a 6 Getting SALES Gawd wile At dae Sus Biel eae ce aie oe Cree es 6 2 1 DEPENGEN SUES esis tri keina E a o E ee we e a 6 2 1 1 Baseline requirementS ooooooooooo eee ee eee 6 2 1 2 Python on non windows platforms 6 Field PYtAON ON MIEDOS ds AA a hehe es AE Sars Ge 6 2 1 4 Obtaaning INUOMPy and SCALP Y ias aliada e seas 7 2 2 Installation set ais e ale SAN Ghd tte thd oe E 7 2 2 1 Installing Python NumPy and SciPy 7 2 202 Unpack the UNLCOR archives diene fared it ole eee See eee 7 Zar sta UNG COR eater ee Sion ns atte gate he onda te Rod cinerea a et Woks 7 2 2 4 Optional Python extension modules 8 230 Example LUNA
6. agmentation under future climate or landuse change 1 2 How does UNICOR work UNICOR simulator uses a modified Dijkstra s algorithm Dijkstra 1959 to solve the single source shortest path problem from every specified species location on a landscape to every other specified species location Figure 1 provides a step by step conceptual workflow UNICOR requires two input files as the first step 1 a landscape resistance surface and 2 point locations for each population or individual s location see section 3 for program input Prior to running UNICOR users must create a resistance surface where each cell value pixel represents the unit cost of crossing each location Pixels are given weights or resistance values reflecting the presumed influence of each variable to movement or connectivity of the species in question e g Dunning et al 1992 Cushman et al 2006 Spear et al 2010 Resistance surfaces could be parameterized to reflect different costs to movement associated with vegetation types UNICOR user manual 4 elevation slope or other landscape features UNICOR Conceptual Workflow DEFINE PROBLEM Step 1 Assemble Step 2 Set Input Geospatial Inputs Parameters Figure 1 UNICOR conceptual workflow diagram Steps 1 and 2 define the inputs and problem Steps 3 and 4 execute the progr
7. am Step 5 provides synthesis and post processing Point locations define starting and ending nodes of paths connecting pairs of individuals The points must be referenced on the landscape resistance surface with any user specified placement pattern e g uniform random or placement in habitat suitability and density From graph theory and network analysis we can then represent the landscape resistance surface as a graph with nodes and edges Diestel 2010 Every pixel is considered to be a node The graph edges which represent possible movement paths between each node are weighted by the resistance value of the cell times the distance to the next pixel center times the distance to the next pixel center which gives the total edge length in terms of raster cell units cost istance Dijkstra s algorithm is modified in the UNICOR code to find all shortest paths to all destination nodes associated with the same starting node This provides a substantial boost in computational efficiency where all pairwise combinations are found for the same starting node before clearing the search space from memory All paths found are optimal paths of movement computed for every paired combination of starting and ending nodes The combination of these shortest paths create a path density map which is also a connectivity graph In essence this approach becomes a large graph proble
8. ame directory 2 The included rip file specifies the parameters that can be changed and used in the sample UNICOR run Open example rip in your editor of choice e g notepad or wordpad 3 This file is the stanza format following RipMgr documentation All signs are comments followed by variable names with a tab to the parameter specified The parameter can be changed for running UNICOR but downloaded parameters will run as is See section 3 for more details on each parameter along with its dependency 4 Start the program with a graphical interface or at the command line For example if you downloaded Python 2 6 x from www python org then you are provided with a graphical interface IDLE n Windows you can find IDLE from your Start menu gt All Programs gt Python 2 6 gt DLE Python GUI Alternatively if you use Python from the command line then open a terminal window and change your shell directory to the UNICOR home directory 5 Run the program There are a number of ways to run this program For example if you are using a command shell you can run the program by typing python UNICOR py example rip 6 Check for successful simulation run completion The program will provide a log file in your UNICOR home directory Once completed UNICOR user manual 9 res output files will be created in UNICOR home
9. cography Submitted 5 5 Disclaimer The software is in the public domain and the recipient may not assert any proprietary rights thereto nor represent it to anyone as other UNICOR but not than a University of Montana produced program version 1 x is provided as is without warranty of any kind including limited to the implied warranties of merchantability and fitness for a particular purpose The user assumes all responsibility for the the accuracy and suitability of this program for a specific application In no event will the authors or the University be liable for any damages including lost profits lost savings or other incidental or consequential damages arising from the use of or the inability to use this program We strongly urge you to read the entire documentation before ever running UNICOR We wish to remind users that we are not in the commercial software marketing business We are scientists who F recognized the need for a tool like UNICOR to assist us in our research on landscape ecology issues Therefore we do not wish to spend a great deal of time consulting on trivial matters concerning the use of UNICOR However we do recognize an obligation to provide E some level o by lways interested in learning about how others have applied e cological investigation and management application The
10. command shell s paths or that python is not installed In the first case ask an administrator to add it to your command paths If your shell locates and loads python type import numpy Similarly type import scipy If python does not complain that there are no such modules all is well The following instructions assume Python NumPy and SciPy are not yet available on your computer if they are skip to section 2 2 2 First run the Python executable installer you ve chosen either from www python org or ActiveState accepting defaults for the installation directory On Windows this will typically place the executables and libraries in c Python2 6 bin and the site packages package tree for user installed Python modules in c Python2 6 lib site packages If you are installing it on a network on which you do not have administrative privileges you may need to ask a system administrator to install python and the NumPy and SciPy packages in their default locations Next install NumPy and SciPy using the supplied executable superpack installer or visiting http www scipy org Download This will install NumPy and SciPy in your Python site packages directory 2 2 2 Unpack the UNICOR Archive Navigate to the directory on your PC where you wish to install UNICOR and unpack the supplied zip archive file using a free archive tool like 7Zip
11. ction such as a Gaussian bell Pinto and Keitt 2008 is followed by all kely to realistically represent the bene or to smooth output paths using a probability ve icial to consider either curve Cushman et al 2008 UNICOR implements the latter and allows for the application of a variety of smoothing functions referred to as kernel density functions Gaussian biweight triweight density estimations Li amp Racine 2007 by the program show the cumulative de by kernel density estimations following a distribution around frequency of Through batch capability affected by landscape change and scenarios e g Cushman et al g maintaining network connectivity habitat fragmentation outputs of graph theory metrics radius etc Hagberg et al 2008 d E RAGSTATS McGarigal et al 2002 The program is written in Python 2 6 module plug in modular development name value pairs in a stanza oriented text file format Epanechnikov and cosine functions can be used for the kernel see Silverman users may specify al thresholds to assess how scale dependency of dispersal 2010a include paths among habitat patches that can be xpected species movement routes and can uidance on identifying corridors that are likely critical Quantification of changes to and corridor connectivity e g and connectivity outputs that can irectly input into pop
12. currently only available in the UNICOR user manual 10 Linux operating system For parallel computing specify the number of processors that are used in a simulation 3 5 Kernel density estimation Dijkstra s base algorithm assumes the optimal is followed by all individuals However this is unlikely to realistically represent the behavior of organisms Thus it is beneficial to consider either multiple low cost paths or to smooth output paths using a probability density function such as a Gaussian bell UNICOR implements the latter and allows for the application of a variety of smoothing unctions referred to as kernel density functions Gaussian panechnikov uniform triangle biweight triweight and cosine unctions can be used for the kernel density The outputs that are produced by the program show the cumulative density of optimal paths buffered by kernel density estimations following a distribution around E e frequency of common connections h m E 4 Output Outputs include paths among habitat patches that can be used to display expected species movement routes and can provide managers with visual guidance on identifying corridors that are likely critical for maintaining network connectivity Quantification of changes to habitat area fragmentation and corridor connectivity is enabled
13. directory 3 Input See Table 1 for UNICOR inputs and outputs 3 1 Resistance grid Prior to running UNICOR users must create a resistance surface where each cell value pixel represents the unit cost of crossing each location Pixels are given weights or resistance values reflecting the presumed influence of each variable to movement or connectivity of the species in question Resistance surfaces could be parameterized to reflect different costs to movement associated with vegetation types elevation slope or other landscape The filename for the resistance surface must be in an ascii format with header file any file extension is acceptable must be space delimited The example simulation runs are small _test rsg and medium test rsg 3 2 XY locations Point locations define starting and ending nodes of paths connecting pairs of individuals The points must be referenced on the landscape resistance surface with any user specified placement pattern e g uniform random or placement in habitat suitability and density The filename for the individuals with x y locations can have any file extension but must be comma delimited In addition points must fall in unique pixel of the resistance grid i e the point spacing should be greater than the square root of your resistance grid resolution size If you have overlapping points the program will display the following error messa
14. ge There are overlapping points around x and y please check for points that are too close together This point will not be included in the run The example simulation runs are small test 10pts xy and medium test 100pts xy 3 3 Thresholding To reflect species specific differences in dispersal abilities users can specify connectivity thresholds These connectivity thresholds are expressed as the maximum path length for a species given its dispersal ability This enables UNICOR to realistically reflect the biological dispersal abilities of a particular species Users can specify the maximum dispersal distance based either on cumulative cost distance or Euclidean distance To get all path lengths set the Edge Distance to max 3 4 Number of processors In essence this approach becomes a large graph problem for the conservation biology problems faced today In analyses involving large numbers of individuals across a large and fine grained environment computational time becomes intractable However parallel processing allows for efficient use of increasingly ubiquitous modern multi core processors Dijkstra s breadth first search algorithm is ideal for running in parallel for sets of source and destination points because pairwise distances can be calculated independently We have implemented parallel processing in UNICOR using the multiprocessing module from Python version 2 6 and is
15. ice and Visualization John Wiley and Sons New York Schwartz M K et al 2009 Wolverine gene flow across a narrow climatic niche Ecology 90 3222 3232 Silverman B W 1986 Chapter 3 In Density Estimation for Statistics and Data Analysis Chapman and Hall New York Spear S F et al 2010 Use of resistance surfaces for landscape genetic studies Considerations for parameterization and analysis Molecular Ecology in press Urban D and Keitt T 2001 Landscape connectivity A graph theoretic perspective Ecology 82 1205 1218 Watts K et al 2010 Targeting and evaluating biodiversity conservation action within fragmented landscapes an approach based on generic focal species and least cost networks Landscape Ecology 25 1305 1318 7 Acknowledgements This research was supported in part by funds provided by the Rocky Mountain Research Station Forest Service U S Department of Agriculture and by the National Science Foundation grant DGE 0504628
16. ics Output delimited graphmetrics 5 5 1 F General issues The program is including manual l instructions How to obtain UNICOR reeware and can be downloaded at http cel dbs umt edu software UNICOR with information for users FAQ publications and developer involvement 5 2 Debugging and troubleshooting ongoing research For help with installation problems please check first for postings at our web site Otherwis pleas me at erin landguth mso umt edu report problems including any bugs to UNICOR user manual 12 5 3 UNICOR limitations The following is a list of the current as we know of limitations with UNICOR 1 The resistance surface is in ASC format header file with 6 lines of information and space delimited 2 The point locations have a header row and file is comma delimited 3 Point locations must fall inside the resistance grid extent code will run when points lie outside of grid but no paths will be calculated 5 4 How to cite UNICOR The This program was developed by Erin Landguth Brian Hand and Joe Glassy GUI development was done by Mike Jacobi Ross Carlson assisted with graphics data set and website creation The reference to cite is Landguth EL Hand BK Glassy JM Cushman SA Carlson RT 2011 UNICOR A species connectivity and corridor network simulator E
17. m for the applied landscape connectivity assessments UNICOR user manual 5 In analyses involving large numbers of individuals across a large and fine grained environment computational However parallel processing time becomes intractable processing allows for efficient use of increasingly ubiquitous search algorithm is ideal modern multi core processors for running in parallel Dijkstra s breadth first for sets of source and destination points because pairwise distances can be calculated independently processing in UNICOR is current We have implemented parallel processing in UNICOR using the multiprocessing module from Python version 2 6 y only impl Parallel mented under the Linux operating system To reflect species specific differences can specify connectivity thresholds in dispersal abilities users These connectivity thresholds are expressed as the maximum path length for a species given its dispersal ability This enables biological dispersal abilities o specify the maximum dispersal distance or Euclidean distance E Dijkstra s base algorithm assumes the optimal UNICOR to realistically reflect the a particular species distanc Users can ither on cumulative cost based individuals However this is unli behavior of organisms Thus it is multiple low cost paths density fun
18. n older Python interpreter pre v2 4 with a newer one v 2 6 x on a Linux or MAC OS X machine can be tricky so ask a System Administrator for help if you re not sure which packages depend on the current Python installed 2 1 3 Python on Windows Windows 7 XP 2000 Server does not come with Python installed so follow the instructions below to obtain and install Python ona computer running the Windows operating system Get a windows installation of the base Python installation current v 2 6 x at http www python org download releases UNICOR user manual 7 2 1 4 Obtaining NumPy and SciPy We recommend using the superpack Windows installer available from the SourceForge website http sourceforge net project Note that more complete information for NumPy is available at www scipy org where the SciPy module is also presented Another source is http www enthought com products epd php for a free academic and educational usage in a single downloadable installer that has everything and then some Numpy Scipy Matplotlib and 70 modules for python 2 2 Installation 2 2 1 Install Python NumPy and SciPy Make sure that Python and NumPy are installed and available to you You can test this by typing python at a command window If python is available you ll get the python prompt gt gt gt If it is nota recognized command it means either that python is installed but is not in your
19. ransform Linear Specify the scaling used for the function resistance kernel linear scale kernel height to be between 0 and 1 using a linear function Const kernal Tru Keep kernel volume constant for each vol point and 1 Number of 8 For parallel computing the number of Linux Processors processors that are used in a simulation KDE Function Gaussian The probability distribution used to SciPy calculate the kernel density buffer Gaussian Epanechnikov Uniform Triangle Biweight Triweight Cosine KDE GridSize 2 The kernel buffer window used to SciPy calculate the buffered maps Number of Levels 3 The number of categories used to display SciPy the kernel density buffer map Output Default Description Dependency Example Input Save Path Output TRUE The surface of paths in ascii format with header file space delimited addpaths Save Individual TRUE The list of individual path values and Paths Output length from every point to every other point comma delimited paths Save Cost TRUE The resistance distance matrix of all Distance Matrix the source destination connection Output lengths comma delimited cdmatrix Save KDE Output TRUE The surface of kernel buffered paths in ascii format with header file space delimited kdepaths Save Levels TRUE The categorical surface created from the Output KDE output in ascii format with header file space delimited levels Save Graph FALSE Path graph theory metrics comma NetworkX Metr
20. refore we ncourage you to contact us and describe your application after using information support Of course we welcome and ncourage your criticisms and suggestions about the program at all imes We will welcome questions about how to run UNICOR or interpret e output only after you have read the entire documentation nly fair and will eliminate many trivial questions Finall Yr UN This is we are COR in Gaooov odcecto Z ICOR UNICOR user manual 13 We hope that UNICOR is of great assistance in your work and we look forward to hearing about your applications 6 References Bunn A G Urban D L and Keitt T H 2000 Landscape connectivity A conservation application of graph theory Journal of Environmental Management 59 265 278 Compton B et al 2007 A resistant kernel model of connectivity for vernal pool breeding amphibians Conservation Biology 21 788 799 Cushman S A et al 2006 Gene flow in complex landscapes testing multiple hypotheses with casual modeling The American Naturalist 168 486 499 Cushman SA McKelvey K S and Schwartz M K 2009 Use of empirically derived source destination models to map regional conservation corridors Conservation Biology 23 368 376 Cushman S A Chase M J and Griffin C 2010a Mapping landscape resistance to identify corridors and barriers for elephant movement in southern Africa In Cushman S A and Huettman F eds Spatial comple
21. rithm to individual based simulations Outputs can be used to designate movement corridors identify isolated populations quantify effects of climate and management changes on population connectivity and prioritize conservation plans to maintain population connectivity The key features include a driver module framework connectivity mapping with thresholding and buffering kernel resistance thresholoding and calculation of graph theory metrics Through parallel processing computational efficiency is greatly improved allowing analyses of large extents grid dimensions of thousands and large populations individuals in the thousands 1 1 What can UNICOR do UNICOR is intended for use by land managers as well as the research community and will be a valuable tool in applied conservation biology It provides new functionality to increase understanding of species connectivity in current and future landscapes This in turn provides invaluable ability to quantitatively compare spatially explicit conservation and restoration scenarios and prioritize actions to have the largest cumulative effects on population connectivity The results can be used to designate sites as potential source or sink populations and identify corridors and barriers Simulations could address prioritizing areas of greatest concern effects of climate change on wildlife populations or habitat fr
22. ttern Analysis Program for Categorical Maps Computer software program produced by the authors at the University of Massachusetts Amherst Available at the following web site UNICOR user manual 14 http www umass edu landeco research fragstats fragstats html McRae B H and Beier P 2007 Circuit theory predicts gene flow in plant and animal populations Proceedings of the National Academy of Science USA 104 19885 19890 McRae B H et al 2008 A multi model framework for simulating wildlife population response to land use and climate change Ecological Modelling 219 77 91 Li Q and Racine J S 2007 Nonparametric Econometrics Theory and Practice Princeton University Press Opdam P and Wascher D 2003 Climate change meets habitat fragmentation linking landscape and biogeographical scale levels in research and conservation Biological Conservation 117 285 297 Pinto N and Keitt T H 2009 Beyond the least cost path evaluating corridor robustness using a graph theoretic approach Landscape Ecology 24 253 266 Riitters K et al 2000 Global scale patterns of forest fragmentation Conservation Ecology 4 online URL http www consecol org vol4 iss2 art3 Sawyer H et al 2009 Identifying and prioritizing ungulate migration routes for landscape level conservation Ecological Applictions 19 2016 2025 Scott D W 1992 Chapter 6 In Multivariate Density Estimation Theory Pract
23. ular landscape pattern analysis programs uniform triangle The outputs that are produced nsity of optimal paths buffered 1986 Scott 1992 common connections F ternative connectivity ability will be fragmentation under a range of Watts et al 2010 Outputs used to display provide managers with visual for is enabled through number of nodes density e g UNICOR is built on a driver docking architecture that allows for ease of future The program s input parameters are organized as F The inputs UNICOR user manual 6 gl are parsed using the RipMgr package a exible symbol table manager for science models that includes special parsing capabilities Glassy 2010 UNICOR has been debugged as carefully as possible by testing all combination of simulation options The program is freeware and can be downloaded at http cel dbs umt edu software UNICOR 1 3 What s new in Version 2 0 Similary to connectivity thresholding we have added a resistant kernel approach for predicting habitat connectivity and corridor paths for the given resistance surface Resistant kernel connectivity modeling has a number of advantages as a robust approach to assessing population connectivity for multiple wildlife species First unlike most corridor prediction efforts it is spatially synoptic and provides prediction and mapping
24. xity informatics and wildlife conservation Springer Tokyo pp 349 368 Cushman S A Compton B W and McGarigal K 2010b Habitat fragmentation effects depend on complex interactions between population size and dispersal ability Modeling influences of roads agriculture and residential development across a range of life history characteristics In Cushman S A and Huettman F eds Spatial complexity informatics and wildlife conservation Springer Tokyo pp 369 387 Dale V H et al 2001 Climate change and forest disturbances BioScience Ej 723134 Diestel R 2010 Graph Theory Springer Verlag Heidelberg Fourth Edition Dijkstra E W 1959 A note on two problems in connexion with graphs Numerische Mathematik 1 269 271 Dunning J B Danielson B J and Pulliam H R 1992 Ecological processes that affect populations in complex landscapes OIKOS 65 169 175 Fall A et al 2007 Spatial graphs principles and applications for habitat connectivity Ecosystems 10 448 461 FAO 2006 Global Forest Resources Assessment 2005 Main report Progress towards sustainable forest management FAO Forestry Paper 147 Rome p 320 Hagberg AA et al 2008 Exploring network structure dynamics and function using NetworkX In Varoquaux G et al eds Proceedings of the yee Python in Science Conference SciPy2008 Pasadena CA USA pp 11 15 McGarigal K et al 2002 FRAGSTATS Spatial Pa
Download Pdf Manuals
Related Search