Marxan Tutorial - Peter Arcese Lab
Contents
1. As you can probably imagine testing multiple values increase the computational time as the scenario will have to be run three times in its entirety totaling 300 runs if you keep Number of repeats at the default However it may be useful to test and compare the sensitivity of your study area to changing key variables 11 2 2 4 Connectivity Connectivity Minimum value 0 Maximum value 0 Number of values 1 hje increase by orders of magnitide of min This section will only be included if you have selected Yes to Include connectivity in the analysis in Global Parameters See section 2 2 1 If you have done so then you can change the Minimum value to the boundary length modifier BLM of your choice You may want to start by setting it to 1 As with the Species Penalty Factor section 2 2 3 you can test a range of values by setting the Number of Values slider to a number greater than one and keeping the box beneath checked 2 2 5 Number of Iterations Number of iterations Minimum value 100000 Maximum value 10000000 Number of values 1 SG increase by orders of magnitide of min This section specifies how many iterations Marxan will test per run Each iteration is a specific configuration of planning units so the more iterations that are tested the greater the likelihood that Marxan will a configuration with a low objective function score The Marxan Good Practices Guide s2. modifier the connectivity value the number after that for its value stands for number of iterations the number after that for its value 2 3 2 Summary Tables The tool presents two key table types to summarize the results of the Marxan session and these can be viewed in the lower half of the results section First the overall summary table for the whole session which appears automatically under the Summary tab This summary table provides you with the average or count values for performance in cost connectivity and meeting targets across all runs with the same parameters The second type of table summarizes the performance of the solution from each run If you are testing multiple values of SPF BLM or iterations you will get one table for each set of parameters This table is in the same format as the summary table and can be viewed under the tab labelled with the run name ex S3_B1_11e 05 Note If you don t see a scroll bar use the arrow keys to scroll left and right on the table 13 Table 2 Summary Table Headers and explanation Runname The name given to a set of runs with the same parameters Score The average objective function score across all solutions Cost The average cost across all solutions Measured in units of area assessed value or Pus human score see section 2 2 1 The average number of planning units or land parcels selected in solutions B_length Penalty The average total b
3. nature of your project cost could be calculated as the area of planning units the costs of ongoing management opportunity costs of displaced commercial activities costs to industry tourism and recreation from displaced activities or acquisition cost The lower the cost of a unit the lower the score will be and the more likely it is that the planning unit will be included in the solution Marxan then summarizes the cost of all of the selected planning units and this is incorporated into the score Connectivity The boundary length of the reserve system is way of quantifying the connectivity of a configuration of planning units It is a combination of the total length of the edges of the selected planning units and the weight that you choose to give to this value This weighting is known as the boundary length modifier BLM Essentially If you choose to place importance on the boundary length i e you set the boundary length modifier to a value greater than 0 then configurations with many small and isolated patches will have higher scores Marxan works to find the solution with the lowest score so your reserve system will have a more clumped distribution For this tutorial it will be important to understand the boundary length modifier and the consequences of constraining boundary length on the near optimal solution that Marxan produces Skip ahead to section 2 2 4 for instructions on the practical use of the BLM See section 2 5 for exa
4. of your planning decisions It is not essential to understand how the algorithm works to use Marxan but those wishing to gain a better technical understanding should consult Marxan s User Manual 1 2 How does Marxan work 1 2 1 In a nutshell After you specify your conservation targets and objectives e g area to be conserved Marxan will start by assigning scores to planning unit configurations in the sample space This score is based on that particular configuration s ability to meet the conservation objectives while minimizing the cost the lower the score the better Marxan will then compare a huge number of configurations it would be virtually impossible to compare them all and find the one with the lowest overall score This is your solution You can repeat the process as many times as you want the more times you do the more confident you can be that you have found a near optimal solution 1 2 2 A little more about the scoring of planning units The score that Marxan assigns to each planning unit configuration is based on a mathematical formula called the objective function The complete formula is Score of the configuration being tested gt Cost Boundary Length Modifier x Boundary Length of the reserve system gt Species Penalty Factor x Penalty incurred for unmet targets Cost By costs we are referring to the cost of including that particular set of planning units in a configuration Depending on the
5. old forest communities Calculated using the formula B 2 OF SAV OF SAV Schuster et al 2014 23 Appendix F Garry Oak Ecosystem Plant Native Species Richness Map N SEZA baked Nes GOE Native S Species Richness EE 1 51 2 90 Be 2 91 3 78 IY 3 79 4 49 Mi 4 50 5 20 MM 5 21 5 99 MW 6 00 6 91 MN 6 92 7 96 B 7 97 9 21 MM 9 22 10 82 MM 10 83 13 70 Roads EAF hiss CA PIET T tA y x E g 1 O age Predicted species richness of native Garry Oak and maritime meadow plants MacDougall et a 2006 Bennett and Arcese 2013 Boag 2014 Appendix G Example Maps Locking In vs Not Locking In Parks 0 20 HB 21 40 He 41 o at Mc 20 bee cord Ts HB Best solution HM 00 a A DY Protected Areas 20 Kilometers 20 Kilometers li a a Aa AT Ce a a NS ASS a CP Selection frequency map and best solution map for a Marxan simulation on the Capital Regional District with the following specs Parks Locked in Yes 100 Runs BLM set at 10 SPF set at 3 100 000 iterations 0 20 EE 21 40 HE 41 co oF Hc 20 ie aw HEB Best solution HM 100 b i te es DY Protected Areas 20 Kilometers 20 Kilometers ee sae Ee SS Se Da Pa a S ee A SY Ca CS Selection frequency map and best solution map for a Marxan simulation on the Capital Regional District with the following specs Parks Locked in No 100 Runs BLM set at 10 SPF set at 3 100 000 iterat
6. values at default levels Solution Score terations Cost SPF Connectivity Solution Score Iterations Download S3_B1_Ile 05 3 B10_11e 05 53_B100_11e 05 te c i _ O Ww t O g x 5 QO 105 110 115 Solution Score of best solution Like the cost plot this plot shows the cumulative number of solutions that have a value greater than that of the best solution In this case the value is the overall objective function score expressed as percentage larger than the best solution score Again the changing parameters can be seen to change the shape of the curve 2 4 Downloading and Viewing MARXAN results in ArcMap The table that will be the most useful for visualizing your results is the Summary Attribute Table located under the Download tab This is an attribute table with the summed solutions and best solution that you can join to the CDFCP shapefile The column header with a _B after the run name is the scenarios best overall solution expressed as O or 1 depending on whether the planning unit was chosen The column header with no _B after the run name gives the selection frequencies of planning units out of 100 runs or however many runs were requested The other table is the Individual Runs Attribute Table This is similar to the previous output but instead of including the selection frequency of parcels or the best run the solutions for all 100 runs of a scenario ar
7. 012 Using bird species community occurrence to prioritize forests for old growth restoration Ecography 35 1 9 Schuster R T G Martin P Arcese 2014 Bird community conservation and carbon offsets in Western North America PLOS ONE 9 1 9 Seely B 2012 Evaluation of carbon storage within forests in the Coastal Douglas Fir zone 14 pages 17 4 Additional Resources Marxan User Manual Ball I R and H P Possingham 2000 MARXAN V1 8 2 Marine Reserve Design Using Spatiall Explicit Annealing a Manual Marxan Good Practices Handbook Ardron H P Possingham and C J Klein Eds Version 2 2010 Marxan good practices handbook University of Queensland St Lucia Queensland Australia and Pacific Marine Analysis and Research Association Vancouver British Columbia Canada Another great resource is the tutorial available on the Marxan website which goes through much of the same material in greater detail htto www uq edu au marxan tutorial toc htm 18 5 Appendices Old Forest Community Score E 0 02 0 14 E 0 15 0 22 E 0 23 0 31 WH 0 32 0 40 E 0 41 0 49 E 0 50 0 57 E 0 58 0 65 E 0 66 0 72 Me 0 73 0 78 E 0 79 0 90 Roads Composite distribution map based on probability of occurrence of birds typically associated with old forest habitat Schuster and Arcese 2014 Appendix B Savannah Community Occurrence Map Savannah Community Score E 0 03 0 14 E 0 15 0 2
8. 2 2 Therefore you may want to test the usefullness of this function by running Marxan with and without a road density constraint Parcel Size Marxan will consider all land parcels in a solution including small and isolated forest patches that may be at considerable risk from encroaching development Unless there are significant conservation features on a property that you are interested in consider placing a constraint on the minimum allowable size Users of the tool found that a minimum size of 2 hectares was a good cut off for identifying securable properties with high conservation value Agriculture density Agriculture includes cultivated fields orchards vinyards and golf courses and is measured as square kilometers of agriculture per square kilometer Values were calculated using Terrestrial Ecosystem Mapping TEM data p Old Forest Community 012 3 4 5 6 Agriculture sqrt ha km sq Figure 3 Graph showing relationship between agriculture and bird species typically associated with old forest communities stands 280 years N 1248 R 0 42 Parcels with more than 3 hectares of agriculture per km demonstrate an about 50 decline in the probability of encountering an old forest associated community Excluding properties with agricultural density ie gt 3ha km may therefore help fine tune your solution 2 2 2 Protection Targets Protection Targets Global Protection Target 17 Set Individual Targets Global Targ
9. 4 E 0 25 0 35 HH 0 36 0 43 WN 0 44 0 50 BE 0 51 0 56 E 0 57 0 61 E 0 62 0 67 E 0 68 0 73 E 0 74 0 92 Composite distribution map based on probability of occurrence of birds typically associated with savannah habitat Schuster and Arcese 2014 Appendix C Wetland Community Occurrence Map N EES y ee Wetland i iG Community Score 0 00 0 07 0 08 0 14 EE 0 15 0 20 E 0 21 0 25 EE 0 26 0 30 E 0 31 0 36 s 1 E 0 37 0 43 i i MM 0 44 0 52 A MM 053 0 62 Se a S V HM 0 63 0 96 ht We a iN PILET E 12 erik 4 Composite distribution map based on probability of occurrence of birds typically associated with wetland and riparian habitats Schuster and Arcese unpublished Appendix D Human Commensal Birds Community Occurrence Map Human Commensal Community Score 0 00 0 09 Be 0 10 0 14 E 0 15 0 19 E 0 20 0 25 HH 0 26 0 32 GE 0 33 0 40 E 0 41 0 49 MH 0 50 0 59 HM 0 60 0 69 E 0 70 0 85 Composite distribution map based on probability of occurrence of birds typically associated with urban and rural human landscapes Schuster and Arcese unpublished 22 Appendix E Old Forest and Savannah Beta Diver sity Map Beta Diversity Score Ma 0 04 0 14 Me 0 15 0 23 Ge 0 24 0 31 MH 0 32 0 39 WN 0 40 0 47 MH 0 48 0 54 MM 0 55 0 59 Hi 0 60 0 65 E 0 66 0 70 E 0 71 0 381 Roads Both savannah and
10. Marxan Tutorial 1 B ckeroun d IMI ONMALION esene a A 2 ELWY Maran i a E E E S 2 1 2 What act ally 1S Marxah en A A EEE E E E 3 T2 Howdoes MIxamM WOK minnes er a sume nenee ticles sles ex catndyedauksamecal wees ens 3 kz Me Meh eV NVM E ence reeled eats decane cer raat ccc te ceed eaters intent ee toc ncaa TIE A E N 3 1 2 2 A little more about the scoring Of planning units cccccccessecccceseccceeeseceseeececseeeceeeeeecessuneses 3 t3 What Information does Marxan USC xiacicicccsiasbsteleciccedh dab vdeleced ceedaasbenetecie in taatvbelets dan neceadebatetemneasiades 4 Usine teM er ICE aare tae caved en a a a ea wide a 5 PGE CTI Se tarne eonan a a a a nnn a se amiaeennien 5 22 Manipulatiic KEV Varla DIES sea a a N oetecateaunaaeaeueeeees 6 DDN Gopal Paramete S erna aE A E E A ENE EENE 6 Zed PD PrODErV EXCIUSION eonna a N a E 8 2 2 A Protection Targets nn a T T a an I N N 9 2 23 Species Penalty Factor are aae N A E teat as pe testes hodelad 11 22 A CONMEGCUIVILY esin a a N a e a E 12 ZP NUMDEr orter ONS rn N E T S 12 2 3 Running Marxan and Interpreting the Res lts inisini E a 13 DSi d SC CNarIO WANG reds E damaet ened ue tenon EE he cciMat AA E E ede E dead es 13 2 D2 SUMMARY VAIO OS aana E te waite a sw aidels evant aden E N 13 DBS OUR PUT DIOL a aana acs acts aaa a ees ea ee nee 15 2 4 Downloading and Viewing MARXAN results in ArcMap ccccssseccceesecccceeseceeeeececeeeceeseuseceseeneces 16 RETORCTIC CS e
11. a ar E enous anu aigestaws a S 17 AAGITIONGlsRESOUN COS ccscthadornasarccanniasenpsteasennndandeunasosasanan adenine vbabeunstuasaune accannsandemeeoubeunetoasemns abeewennteness 18 ADDON CICES ccc neu A E esuamedans sete a E a eee ete 19 Appendix A Old Forest Community Occurrence MaP cccccsssecccceseccecesccecsesecceeeeecceseuecessuneceeseneeeetas 19 Appendix B Savannah Community Occurrence Map cccsccccssecccssceceeececeuecseeneceeenseeeeeceseueceeeneseeeneeees 20 Appendix C Wetland Community Occurrence Map cssccccssseccccsseccccesececsenecceeeecceseesecesseneceeseneeeeeas 21 Appendix D Human Commensal Birds Community Occurrence Ma p c ssseccccssseceeeesececeeeeceeeeeeeeeeas 22 Appendix E Old Forest and Savannah Beta Diversity Map cccccccsssssseeecccceeeeeessesecceessaueeseeeeeeeeesaaas 23 Appendix F Garry Oak Ecosystem Plant Native Species Richness Map cscccccssseccceesececeeseceeeeeseeeees 24 Appendix G Example Maps Locking In vs Not Locking IN Parks ccccccssseccccesseceeesececeeeeceeeeeeeeeeas 25 Marxan Tutorial Produced for the Coastal Douglas Fir Conservation Partnership htto arcese forestry ubc ca marxan tool 1 Background Information 1 1 Why Marxan Marxan is a problem solving tool that is used to help inform decisions on landscape scale conservation planning As part of a systematic planning process Marxan contributes towards a transparent inclusive
12. and defensible decision making process Historically conservation decision making has often focused on evaluating land parcels opportunistically as they become available for purchase donation or under threat This was often done without a complete understanding of how their acquisition might contribute to targets for biodiversity conservation or the degree to which they are likely to maximize return on the investment of scarce conservation dollars and time Using Marxan to simulate alternative reserve designs should help you to prioritize conservation actions at the landscape level and allow you to specify the targets such as focal or indicator species richness ecosystem representation complementarity and connectivity and to also minimize the overall costs of conservation acquisitions This tutorial will show you how to use Maxan to identify existing gaps in biodiversity protection to identify candidate areas to include in a growing reserve system and to provide decision support based on a clear and repeatable set of conservation targets Figure 1 Maps of Salt Spring Island after running a Marxan simulation of 100 runs with a 17 target for all conservation features boundary length modifier of 1 and species penalty factor of 3 The maps on the left indicates the frequency with which particular parcels were selected of 100 runs with dark blue indicating properties that were nearly or always part of the solution and yellow properties n
13. consider using this option if you want to assess the efficacy of the existing reserve system See Appendix G for example maps Include connectivity in the analysis Specify if you want Marxan to include boundary length in the calculation for the objective function See section 1 2 2 If you select No then the boundary length modifier will be set to O and the boundary length of the configurations being tested will not be considered This means that there will be no consideration for the spatial clumping of planning units If you do want to favour solutions with a more clumped distribution then select Yes You can then select a range of boundary length modifiers to test in the Connectivity section see Section 2 2 4 You might wish to consider this option to avoid selecting isolated land parcels but keep in mind that it could considerably increase the cost of the final solution What cost metric should be used Property size area lt Property size area Assessed land value Human score wre eee of us ape ene z Select the cost metric that best suits the goals of your project Property size uses land area as a proxy for cost and is useful if you are interested in protecting a certain percentage of the land base ie 50 regardless of specific property costs Assessed land value is generated using a combination of cadastral data Integrated Cadastral Information Society of BC and 2014 land value a
14. e included _r001 _r100 in the attribute table This table will likely be less useful to you but you can still choose to join it to the CDFCP shapefile in ArcMap for a more detailed look at the solutions for individual runs To start download the Cadastral Fabric property parcel layer for the CDFCP Also download the Summary attribute table 16 To display Marxan outputs correctly you will have to load the cadastral fabric and then join the results attribute table to that layer This step can be a bit tricky because you are joining a csv file to the cadastral shapefile To do so first create a File Geodatabase in ArcCatalog or the Catalog tab of ArcMap You can do this by browsing to your desired folder location in the Catalog right clicking and selecting New gt File Geodatabase Next browse to the downloaded Summary attribute table from the online tool Right click on that and select Export gt To Geodatabase single A Table to Table tool interface will pop up For the Output Location select the recently created File Geodatabase Next in ArcMap right click the cadastral fabric you added select Joins and Relates gt Join In the first drop down menu select Join attributes from a table In the following list of numbered items choose 1 RS_ID 2 Browse to the Marxan output table in the Geodatabase 3 ID or PUID if you are working with the individual runs table And finally in t
15. erent subset of the total CDFCP dataset To look for solutions across the entire CDFCP area follow the first CDFCP wide link r Capital Regional District EE Cowichan Valley Regional District R Islands Trust Area CDFCP Full Extent Figure 2 Extent of coverage for the subsets available in the Marxan tool An important note about scale The spatial scale which you choose has the potential to greatly alter the solutions that Marxan produces A smaller planning area ex Salt Spring Island has fewer land parcels to choose from and as a result Marxan may be forced to consider parcels with lower conservation value in order to meet its targets Expanding your planning area ex CRD could increase the availability of high quality parcels but this may take the focus away from your area of interest It is important to consider the consequences of your subset choice before you run Marxan You may even consider running the same parameters at different scales and comparing the results 2 2 Manipulating Key Variables Once you ve chosen your data subset you can begin manipulating the parameters that Marxan uses to inform the objective function See section 1 2 2 All manipulations are done within the grey sidebar on the left hand side of the screen Once you run Marxan the results will be displayed on the right When you open the tool the key variables will automatically be set to default settings according to the recommendations con
16. et will be ignored In this section you specify your overall objectives or a series of objectives for the biodiversity features contained within the CDFCP area The tool defaults to 17 which is the global target for terrestrial ecosystem conservation as specified in the 1992 UN Convention on Biological Diversity This means that Marxan will try to incorporate a minimum of 17 of each biodiversity feature in its solutions If you wish to vary the protection targets by feature check the Set Individual Targets box and the menu will expand to show all of the biodiversity features for the CDFCP area Table 1 Current biodiversity feature layers in the CPFCP tool Old Forest Birds Savannah Birds Wetland Birds Human Commensal Birds Avoid Human Birds Bird Beta Diversity Standing Carbon Carbon Sequestration Potential TEM Element Occurrence Garry Oak Plant Species SEI Area A composite distribution map based on probability of occurrence of birds typically associated with old forest habitat Schuster and Arcese 2014 See Appendix A A composite distribution map based on probability of occurrence of birds typically associated with savannah habitat Schuster and Arcese 2014 See Appendix B A composite distribution map based on probability of occurrence of birds typically associated with wetland and riparian habitats Schuster and Arcese unpublished See Appendix C A composite distribution map based on probabilit
17. ever selected This output is often thought of as the portfolio of candidate reserves for acquisition stewardship or owner contact by managers The figure on the right indicates the reserve design Marxan returned as the best solution to the input goals and constraints In this case the conservation plan that achieved the highest overall biodiversity score at the lowest overall cost based here on 2014 BC Assessments 1 2 What actually s Marxan Marxan is a computer application that runs an algorithm on a user defined data set and returns a solution in the form of a table of land parcels These parcels form a near optimal balance of input targets and costs Marxan is capable of analyzing large complex datasets to find near optimal solutions because it uses an algorithm called simulated annealing default setting with others possible which selects properties that maximize progress towards your biodiversity goals while minimizing acquisition or other user controlled costs It s important to note that in its basic form Marxan has no graphical interface it just does the computing We have designed a web based graphical user interface that lets you set Marxan parameters and returns a Spatially linked solution file that you can download to ArcMap and view In this tutorial we will introduce you to our user interface explain how to manipulate key variables and gain an understanding about the application of output files in support
18. he Join Options select Keep only matching records Click OK and the Marxan results will be joined to the cadastral fabric To save this join right click the cadastral fabric again and select Data gt Export Data Choose file name and location press OK and after the export is finished select to Add layer On this new layer right click and select properties In the Symbology tab select Quantities In the Value drop down menu select the run name of your choice and set up a color ramp for display 3 References Bennett J R 2013 Comparison of native and exotic distribution and richness models across scales reveals essential conservation lessons Ecography 36 1 10 Bennett J R and P Arcese 2013 Human influence and classical biogeographic predictors of rare species occurrence Conservation Biology 0 1 5 Boag A E 2014 Spatial models of plant species richness for British Columbia s Garry oak meadow ecosystem Master s thesis University of British Columbia Heilman G E J R Stritthold N C Slosser and D A Dellasala 2002 Forest fragmentation of the conterminous united states assessing forest intactness through road density and spatial characteristics BioScience 52 411 422 MacDougall A S J Boucher R Turkington and G E Bradfield 2006 Patterns of plant invasion along an environmental stress gradient Journal of Vegetation Science 17 47 56 Schuster R and P Arcese 2
19. ions 25
20. lty factor will determine how harshly you penalize this deficit A higher SPF value means that Marxan will place a greater importance in solutions that meet targets versus minimizing cost or boundary length However an SPF set too high will be very restrictive and may not allow Marxan to search the sample space efficiently For more information on this refer to the Marxan User Manual yy lt For each section Species Penalty Factor Connectivity and Iterations there is a slider Number of values By default this is set to 1 which means that only the value from Minimum value will be taken for the Marxan run In other words SPF used in the objective function is simply the minimum value But there is the option to test a range of values for tool calibration This is important if you have new data or add data layers as this will change outcomes of Marxan runs By changing the Number of values slider to a number greater than one you are telling Marxan to test that many different SPF values The values of SPF that will be tested will increase in orders of magnitude of the minimum value as long as the box under the slider remains checked For example if your minimum value for SPF is set to 1 and you set the Number of values slider to 3 then Marxan will have to tests SPF values of 1 1 10 and 1 10 or in other words 1 10 and 100 The Maximum value will place a cap on how high those values can be
21. mples of this parameter in practice Species Penalty Factor The piece of the scoring formula is the penalty incurred for unmet targets This is the sum of the user defined penalty for not meeting the target and the weight that you choose to give this value This weighting is known as the species penalty factor SPF As the user you are in charge of setting how big the penalty should be When a planning unit configuration fails to meet a conservation target e g it does not contain a certain level of richness or a target species then it will receive a penalty and this will increase its score by a magnitude proportional to the size of the SPF As a result it is less likely to represent the final solution Skip ahead to section 2 2 3 for instructions on the practical use of the SPF See section 2 5 for examples of this parameter in practice 1 3 What information does Marxan use In order to work Marxan needs to know your project objectives and study area well and this input data needs to be organized into specific file types The key information it requires is 1 Your project area and a list of all of the planning units contained within it as well as their cost 2 A list of target conservation features species habitats soil types A clearly defined objective or series of objectives ex 30 of all grizzly bear habitat How much of each conservation feature is contained within each planning unit Bw The user interface we use in this t
22. oundary length across all solutions The average penalty determined as the sum of SPF the amount of each feature Shortfall that is missing from meeting the targets across all solutions The average shortfall The amount by which the target s have not been met in the Missing Values solution for a run across all solutions The sum of runs with solutions that did not meet the targets MPM The sum of the Minimum proportion met value Table 3 Run Solutions Table Headers and explanation Run_Number Which of the repeat runs the output refers to Score The overall objective function score for the solution for that run Cost The sum of the cost of each planning unit selected for the solution PU s The number of planning units contained in the solution for that run Connectivity The sum of the planning boundaries that form the perimeter edge of the solution for Penalty that run The sum of SPF the amount of each feature that is missing from meeting the Shortfall targets for the solution for that run The amount by which the target s have not been met in the solution for that run Missing Values expressed as a proportion Whether or not the solution has met the target s MPM The minimum proportion met value or in other words the proportion of the worst achieving feature contained within a solution for a run 14 2 3 3 Output plots Cost SPF Cost SPF Connectivit
23. sions Property exclusions If you don t want to exclude properties simply leave values at 0 Road density km km2 Marxan will only select properties with road densities smaller than cutoff 0 ona Parcel size ha Marxan will only select properties bigger than cutoff 0 EEE Agriculture density km2 km2 Marxan will only select properties with agricultural densities smaller than cutoff This section allows you be more specific about the types of land parcels you want included in the solution Leaving any of the sliders at O automatically removes these factors from Marxan analysis Road density Measured as kilometers of paved road per square kilometer and calculated for each land parcel in the CDFCP using Terrain Resource Information Management TRIM data p Old Forest Community Roads sqrt kn km sq Figure 2 Graph showing relationship between road density and bird species typically associated with old forest communities stands 280 years N 1248 R 0 42 The probability of encountering an old forest associated community begins to decline at road densities greater 1 km km Figure 2 Excluding properties with high road density ie gt 1km km may therefore help fine tune your solution Note road density is included in the predictive models for native species cover so setting protection targets for native birds and plants is already indirectly placing a constraint on the road density see Section 2
24. ssessments BC Assessment Agency This metric is generally the most easily translated to acquisition cost and is useful for projects with more constrained budgets Human score is based on a weighting of expert scores for urban and rural areas As this metric identifies human impact rather than monetary value only select it if you wish to focus on biodiversity value and disregard acquisition cost Number of repeats Tell Marxan how many runs you would like it to complete and as a result how many solutions it will produce in the output Because large complex data sets will likely have many near optimal solutions increasing the number of runs will increase the likelihood that you have found the best possible configuration of land parcels the one with the lowest score However this will also increase the time it takes for Marxan to run which can be a major constraint with large datasets and limited time We have set this parameter to a default of 100 runs which should provide adequate repetition to assess which parcels are selected most frequently in solutions as well as provide a good best solution and minimize run time Generate output for individual runs Check this box if you want the option of looking at the solution table for individual runs If you are only interested in the overall best run and summed solution selection frequency outputs then leave this box unchecked to improve processing performance 2 2 2 Property Exclu
25. tained in the Marxan Good Practices Guide 2010 You can choose to keep them in this format or manipulate them to better suit your study objectives We ll go through each section of parameters and explain the options associated with each 2 2 1 Global Parameters This section provides Marxan with basic instructions on how it will run Global parameters How to deal with protected areas Locked in v Include connectivity in the analysis No v What cost metric should be used Property size area v Number of repeats 100 Generate output for individual runs How to deal with protected areas Specify if you want to force Marxan to include existing protected areas and parks in the final solution The two options on the dropdown menu are Locked In and Available If you choose Locked In then every solution Marxan produces will have to include the planning units with a protected status This will be useful if you wish to identify areas to add to an existing reserve system ex If your objective is to increase parks from 6 to 17 For most scenarios Locked In will be the default However many existing parks may be on areas with low conservation value and therefore locking in protection areas may produce solutions with a lower total score than an un constrained solution Choosing Available will allow Marxan to reject any planning unit from the final solution regardless of protection status You may
26. uggests a minimum of 100000 iterations to ensure sufficient exploration of the sample space which is why we have selected this as the default level Increasing this value will increase the computational time but may allow Marxan to find a solution with a lower score As with Species Penalty Factor and Connectivity you can test a number of different values by setting the Number of Values slider to greater than 1 12 2 3 Running Marxan and Interpreting the Results Run Marxan Once you re satisfied that the settings for key variables meet your project requirements you can click the Run Marxan button and a solution for each run will be calculated as well as a best overall solution This computation can take anywhere from 30 seconds to 30 minutes or more depending on how much you are asking it to do Since all of the calculations are done on an external server you don t have to worry about your own computer s computational capacity When the calculations are complete the results section will be populated with plots and tables We ll go through each component and provide a basic explanation For additional information you may want to consult the Marxan User Manual 2 3 1 Scenario name The tool automatically creates a run name for each set of runs with the same parameters which will look something like this S3 B1 I1e 5 S stands for species penalty factor the number after that for its value B stands for boundary length
27. utorial already contains all of the important input information so it is not necessary to understand the specifics how the data is organized However it will be important to understand how to organize your data into input files if you want to run Marxan manually Most of this information is found from external sources Some of it such as boundary length and the conservation features contained within specific planning units are determined using GIS software such as ArcMap or QGis For more information on the input file formats you can refer to Qmarxan an excellent tool for assembling basic Marxan input files 2 Using the Interface 2 1 Getting Started The interface we use is linked to an external server that already contains all of the important input layers for the Coastal Douglas Fir planning area This includes the cadastral fabric cost information existing parks and numerous biodiversity indexes such as old forest birds standing carbon TEM element occurrence etc To connect to the server go to http arcese forestry ubc ca marxan tool in your internet browser The password is Just4Now You will be directed to the following index page Protected Marxan Tool Welcome to the portal of running Marxan tools for the CDFCP Please use the links below to get to the tool version you want fo use CDFCP wide Islands Trust Area Capital Regional District Cowichan Valey Regional District Each hyperlink directs you to a diff
28. y Solution Score Iterations Download 53 Bl Ile 05 53 B10 Ile 05 53 _B100 Ile 05 Cumilative of solutions 110 120 Solution cost of best solution This is the first plot you ll see when Marxan finishes computing It shows the cumulative number of solutions with a cost greater than that of the best solution In this case the reference level for the cost of the best solution is 100 and every other solution will have a cost larger than this adding up until all 100 runs are accounted for The shape of this curve can tell you how much variation in cost there is between solutions which can be important for calibrating SPF values Calibration of the appropriate SPF can be tricky so users should consult the manual for guidance or accept the values included in the tool which are currently set to appropriate values Connectivity Cost SPF Connectivity Solution Score Iterations Download a SPF 3 Nreps 100 Niter le 05 820000 825000 830000 505000 810000 815000 Boundary length 15 The second tab in the results section will show you a plot comparing cost versus boundary length for the BLM values This will be important if you are testing a number of different BLM values See section 2 2 4 Depending on how you set your other parameters it will look something like the above figure which results from setting Minimum Value to 1 Maximum Value to 0 Number of Values to 3 and keeping all other
29. y of occurrence of birds typically associated with urban and rural human landscapes Schuster and Arcese unpublished See Appendix D A composite distribution map based on probability of occurrence of birds that typically avoid urban and rural human landscapes Both savannah and old forest communities Calculated using the formula B 2 OF SAV OF SAV Schuster et al 2014 See Appendix E Total standing carbon per hectare Seely 2012 Predicted carbon sequestration per hectare in the next 20 years Seely 2012 Terrestrial ecosystem map TEM of the Douglas fir Oregon grape community a CDF variant BC Centre for Conservation Data 2014 Predicted native species richness of Garry Oak and maritime meadow plants MacDougall et al 2006 Bennett and Arcese 2013 Boag 2014 See Appendix F Sensitive Ecosystem Inventory Province of BC 2011 Total area target i e Nature Needs Half Note Avoid setting this value higher than the biodiversity targets as Marxan will seek cheap properties to fill the remaining area requirement regardless of biodiversity value 10 2 2 3 Species Penalty Factor Species penalty factor Minimum value Maximum value Number of values el increase by orders of magnitide of min This section specifies the magnitude of penalty for unmet targets For example if only 15 of biodiversity features are included in a configuration when your global target is 17 then the species pena
Download Pdf Manuals
Related Search