Marxan Tutorial for the Coastal Douglas-Fir and
Contents
1. 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 human score see section 2 2 1 Pus The average number of planning units or land parcels selected in solutions B length The average total boundary length across all solutions Penalty The average penalty determined as the sum of SPF the amount of each feature that is missing from meeting the targets across all solutions Shortfall The average shortfall The amount by which the target s have not been met in the solution for a run across all solutions Missing Values 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 that run Penalty The sum of SPF the amount of each feature that is missing from meeting the targets for the solution for that run Shortfall The amount by2. eeeeeeeeneneren enne 23 Appendix E Old Forest and Savannah Beta Diversity Map c cccccccecsssessnsececeeecessessaeseeeesesesessaaeees 24 Appendix F Garry Oak Ecosystem Plant Native Species Richness Map eese 25 Appendix G Example Maps Locking In vs Not Locking In Parks ccccccssssscccececessessaeeeeeeseessessaaeess 26 1 Background Information 1 0 The Coastal Douglas Fir and Associated Ecosystems Conservation Partnership This tutorial was produced as part of an integrative conservation strategy to protect and steward the Coastal Douglas fir biogeoclimatic zone and associated ecosystems in south western British Columbia The Coastal Douglas fir and Associated Ecosystems Conservation Partnership CDFCP is a collaboration of agencies organizations academic institutions and land managers working to balance competing land uses through sound science shared information supportive policies and community education For more information on the CDFCP goals and collaborators refer to the latest version of the Conservation Strategy at http www cdfcp ca index php about the cdfcp conservation strate 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 and defensible decision making process Historically conservation decision ma
3. IN 0 23 0 31 MN 0 32 0 40 E 0 41 0 49 EB 0 50 0 57 IU 0 58 0 65 IN 0 66 0 72 Hl 0 73 0 78 EE 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 0 03 0 14 E 0 15 0 24 IN 025 0 35 E o 36 0 43 E 044 0 50 E 0 51 0 56 E 0 57 0 61 E 0 62 0 67 EN 0 68 0 73 HM 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 Composite distribution map based on probability of occurrence of birds typically associated with wetland and riparian habitats Schuster and Arcese unpublished Wetland Community Score 0 00 0 07 0 08 0 14 I 0 15 020 Em 0 21 0 25 I 0 26 0 30 E 0 31 0 36 EH 0 37 0 43 EN 0 44 0 52 EH 0 53 0 62 HH 0 63 0 96 Roads 22 Appendix D Human Commensal Birds Community Occurrence Map Human Commensal Community Score 0 00 0 09 0 10 0 14 uU 0 45 0 19 E 0 20 0 25 EB 0 26 0 32 E 0 33 0 40 MN 0 41 0 49 IN 0 50 0 59 MMM 0 60 0 69 HMMM 0 70 0 85 Composite distribution map based on probability of occurrence of birds typically ass
4. habitats represents a cost effective approach to conservation planning Which cost metric should be used Property size area 4 Property size area Assessed land value Human score 9 mS ee 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 assessments 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 This is an experimental metric that represents species likely to thrive in human dominated landscapes 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 par
5. 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 212 Using bird species community occurrence to prioritize forests for old growth restoration Ecography 35 1 9 Schuster R T G Martin P Arcese 212 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 18 4 Additional Resources Marxan User Manual Ball I R and H P Possingham 2000 MARXAN V1 8 2 Marine Reserve Design Using Spatially Explicit Annealing a Manual Marxan Good Practices Handbook Ardron J 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 http www ug edu au marxan tutorial toc html 19 5 Appendices Old Forest Community Score J 0 02 0 14 NH 0 15 0 22
6. Welcome to the portal of running Marxan tools for the CDFCP Please use the links below to get to the tool version you want to use CDFCP wide Islands Trust Area FN land excluded Capital Regional District Capital Regional District Vancouver Island only Cowichan Valley Regional District Islands Trust Area FRST495 Each hyperlink directs you to a different subset of the total CDFCP dataset To look for solutions across the entire CDFCP area follow the first CDFCP wide link sm Capital Regional District Il Cowichan valley Regional District A 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 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
7. 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 contained 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 M 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 e g If your objective is to increase parks from 696 to 1796 For most scenarios Locked In will be the default However because existing parks may be located in areas of relatively low con
8. of values increase by orders of magnitide of min This section specifies how many iterations Marxan will test per run Because each iteration returns a specific configuration of planning units when more iterations are obtained the chance that Marxan returns a configuration with a low objective function or near optimal solution increases The Marxan Good Practices Guide suggests a minimum of 100 000 iterations to ensure a full exploration of the sample space which is why we recommend this as the default level Increasing this value increases computation time but may allow Marxan to find a solution with a lower score As with Species Penalty 13 Factor and Connectivity you can test a number of values by setting the Number of Values slider to greater than 1 2 3 Running Marxan and Interpreting the Results Run Marxan Once you re satisfied that the settings for key variables meet your requirements click the Run Marxan button to generate a solution for each run and a best overall solution This computation can take anywhere from 30 seconds to 30 minutes or more depending on what you are asking Marxan to do Since all calculations are done on an external virtual server hosted by the FRBC Chair in Applied Conservation Biolology at UBC Arcese lab 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 now go thr
9. 0 L ER L EE El o D 5 10 20 Kilometers Loa os a pog a a B best Sotien Bl Protected aroas o 5 10 20 Kilometers LL oar oa dp oa r 1 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 iterations 26
10. Marxan Tutorial for the Coastal Douglas Fir and Conservation Partnership Study Area Produced for the Coastal Douglas Fir and Associated Ecosystems Conservation Partnership with funds provided via the FBRC Chair in Applied Conservation Biology Natural Sciences and Engineering Research Council of Canada Real Estate Foundation of British Columbia and The Nature Trust of British Columbia http arcese forestry ubc ca marxan tool Morrell N R Schuster amp P Arcese Department of Forest and Conservation Science University of British Columbia 1 Background Information eren ond pre en mede varied eoe d go veste dia 2 1 0 The Coastal Douglas Fir and Associated Ecosystems Conservation Partnership 2 LT WRyMaFXanD tte ree ipte entrer mn een 2 1 2 What actually is Marxa 2st ttt tea ea ette eet te t bea ue e x dee av e eSI NH eds 3 1 2 How does Marxan WOLK to retener cer en eerte eee Pe eoe E re e aa aaa Eeee deno rtis 4 1 2 11n acnutshell ui epe a Rem EE 4 1 2 2 A little more about the scoring of planning units eese enne 4 1 3 What information does Marxan use ssssssssseseseeeeeeeeee nennen nnne ense en tenent nnne 5 Using the Interface e een ede eun edite npe dientes 6 2 11 Getting Started eU ERR EAUSRUN D EHE a 6 2 2 Manipulating Key Variable Sannir a e a a a Eaa Ei a eaaet ans 7 2 2 1 Global Parameters 4 ze cete te ertet ite te i a aa ie 7 2 2 2 Property Exclu
11. cels 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 set the default to 100 runs which in most cases will provide adequate repetition to assess which parcels are selected most often in solutions 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 This can be good idea in example runs as it allows you to see the overall best run and summed solution selection frequency outputs but you may want to leave this box unchecked to improve processing performance when exploring a range of scenarios 2 2 2 Property Exclusions 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 Parcel size ha Marxan will only select properties bigger than cutoff g 10 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 ea
12. ch land parcel in the CDFCP using Terrain Resource Information Management TRIM data p Old Forest Community Roads sqrt km km sq Figure 3 An empirical relationship of the effect of road density on the occurrence of the old forest bird community in forest stands 280 years of age N 1248 stands r2 0 42 Empirical data based from 700 locations across the CDFCP region indicate that the probability of encountering old forest associated bird communities begins declining at road densities over 1 km km but there is a lot of variation in these observations In part openings associated with some rural roads act as gaps in the forest canopy that can promote the abundance and diversity of old forest species that rely on understory plants Figure 3 Excluding properties with high road density i e 1 3km km may help fine tune your solution away from roaded areas but may also dramatically constrain your solutions in human dominated landscapes Note also that the deleterious effects of roads on native bird and plant communities is also included to some degree in the predictive species maps we used to identify target communities see Appendices Thus by setting protection targets for native birds and plants you are already including constraints on road density to the extent they reduce the value of those biodiversity targets see Section 2 2 2 We therefore suggest you explore the usefulness of this function by running Marxan with and w
13. e distribution map based on probability of occurrence of birds typically associated with wetland and riparian habitats Schuster and Arcese unpublished See Appendix C Human Commensal Birds A composite distribution map based on probability of occurrence of birds typically associated with urban and rural human landscapes Schuster and Arcese unpublished See Appendix D Avoid Human Birds A composite distribution map based on probability of occurrence of birds that typically avoid urban and rural human landscapes Bird Beta Diversity Both savannah and old forest communities Calculated using the formula B 2 OF SAV OF SAV Schuster et al 2014 See Appendix E Standing Carbon Total standing carbon per hectare Seely 2012 Carbon Sequestration Predicted carbon sequestration per hectare in the next 20 years Seely Potential 2012 11 TEM Element Occurrence Terrestrial ecosystem map TEM of the Douglas fir Oregon grape community a CDF variant BC Centre for Conservation Data 2014 Garry Oak Plant Species Predicted native species richness of Garry Oak and maritime meadow plants MacDougall et al 2006 Bennett and Arcese 2013 Boag 2014 See Appendix F SEI Sensitive Ecosystem Inventory Province of BC 2011 Area 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 a
14. esults 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 values at default levels Solution Score lterations Cost SPF Connectivity Solution Score amp Iterations Download 53 B1 l1e 05 3 B10 I1e 05 S3 B100 I1e 05 uw c o g 2 o uv v o o z 5 E 3 o 105 110 115 Solution Score 96 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 uni
15. ing 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 examples of this parameter in practice Last before using a boundary modifier consider existing levels
16. ite crowned sparrows and others Overall your final selection of parcel type should be linked to your specific conservation goals 2 2 2 Protection Targets Protection Targets Global Protection Target 17 Set Individual Targets Global Target will be ignored In this section you specify your overall objectives or a series of objectives for the biodiversity features contained in the CDFCP database The tool defaults to a 17 global target for terrestrial ecosystem conservation as represented in the data set and adopted in the UN Convention on Biological Diversity This means that Marxan will try to include a minimum of 1796 of each biodiversity feature in its solutions We suggest all users to consider carefully all protection targets in consultation with stakeholders With those targets identified you can then check the Set Individual Targets box to reveal a menu showing the biodiversity features currently mapped and available for the CDFCP area These features are listed briefly below Table 1 Current biodiversity feature layers in the CPFCP tool Old Forest Birds A composite distribution map based on probability of occurrence of birds typically associated with old forest habitat Schuster and Arcese 2014 See Appendix A Savannah Birds A composite distribution map based on probability of occurrence of birds typically associated with savannah habitat Schuster and Arcese 2014 See Appendix B Wetland Birds A composit
17. ithout a road density to see its effect on your results Parcel Size Marxan will consider all land parcels in a solution unless you chose to exclude them However small habitats patches do not always represent viable acquisition targets so you can also choose to exclude parcels smaller than a given size using the slider Many beta users of this tool have excluded parcels smaller than 2 hectares from CDFCP solutions based on various assumptions about conservation values and acquisition and stewardship costs in future Agriculture density Agriculture can include cultivated fields orchards vinyards golf courses or greenhouses and is measured in our database as square kilometers of agriculture per square kilometer of land using Terrestrial Ecosystem Mapping TEM data p Old Forest Community 1 0 1 2 3 4 5 G Agriculture sqrt ha km sq Figure 4 An empirical representation of the effect of agriculture on the occurrence of old forest bird communities in stands 280 years N 1248 stands r2 0 42 In parcels with more than 3 hectares of agriculture per km we expect about a 5096 reduction in the probability of encountering the old forest bird community identified by experts as indicating high quality old forest habitat Excluding parcels with a lot of agricultural land may enhance the value of your 10 designs for forest bird species but de emphasize your focus on savanna woodland species such as chipping savanna and wh
18. king 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 1796 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 never selected This output is often thought of as the portfolio of ca
19. ndidate 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 is 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 of your planning decisions It is not essential to understand how the algorithm works
20. ne you are telling Marxan to test that many different SPF connectivity or iteration values In each case the values 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 12 and 100 Maximum value will place a cap on how high those values can be As you can probably imagine testing multiple values can increase the computational time substantially However doing so may be necessary to explore the sensitivity of your outputs to the particular study area and variable combinations chosen 2 2 4 Connectivity Connectivity Minimum value 0 Maximum value 0 Number of values increase by orders of magnitide of min This section will only be relevant if you select Yes under Include connectivity in the analysis in Global Parameters See section 2 2 1 If so then you can change the Minimum value for the boundary length modifier BLM 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 it checked 2 2 5 Number of Iterations Number of iterations Minimum value 100000 Maximum value 10000000 Number
21. ociated with urban and rural human landscapes Schuster and Arcese unpublished 23 Appendix E Old Forest and Savannah Beta Diversity Map N Beta Diversity Score 0 04 0 14 B 0 15 0 23 Ill 0 24 0 31 m 0 32 0 39 E 0 40 0 47 INI 048 0 54 EN 0 55 0 59 IN 0 60 0 65 IN 0 66 0 70 IH 0 71 0 81 Both savannah and old forest communities Calculated using the formula B 2 OF SAV OF SAV Schuster et al 2014 24 Appendix F Garry Oak Ecosystem Plant Native Species Richness Map GOE Native Species Richness 1 51 2 90 2 91 3 78 I 3 79 4 49 Il 4 50 5 20 E 5 21 5 99 MN 6 00 6 91 MM 6 92 7 96 eer EH 797 921 M 0 2 4 IN 9 22 10 82 K m a E 10 83 13 70 Roads et T Us Predicted species richness of native Garry Oak and maritime meadow plants MacDougall et al 2006 Bennett and Arcese 2013 Boag 2014 25 Appendix G Example Maps Locking In vs Not Locking In Parks Ag J 0 20 HH 2 4 L ER Ev El 00 p 5 10 20 Kilometers Loa s s s oas oa a d I 3 B oc oues o 5 10 20 Kilometers LL oca oca poa oa ad E Protected Areas 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 21 4
22. of connectivity allowed via private land management for forestry agriculture recreational or ecosystem values such as carbon storage water purification or pollination services Keeping private land hospitable to native species has the potential to eliminate the need to enforce connectivity via land acquisition 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 Alis
23. ough each of these components to provide a basic explanation below For additional information please 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 I1e405 S stands for species penalty factor the number after that for its value B stands for boundary length 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 l1e405 Note If you don t see a scroll bar use the arrow keys to scroll left and right on the table 14 Table 2 Summary Table Headers and explanation
24. rea requirement regardless of biodiversity value 2 2 3 Species Penalty Factor Species penalty factor Minimum value 3 Maximum value 100 Number of values increase by orders of magnitide of min This section is meant to help you understand the magnitude of penalty for unmet targets For example if only 1596 of biodiversity features are included in a configuration when your global target is 1796 the species penalty factor will determine how severe the penalty for this deficit A higher SPF value causes Marxan to place more importance on solutions that 13 meet targets relative to cost or boundary length However setting SPF too high can be restrictive by preventing Marxan from searching the sample space efficiently For more information on selecting an appropriate SPF refer to the Marxan User Manual For each of Species Penalty Factor Connectivity and Iterations below there is a slider called 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 the SPF used in the objective function will be the minimum value used However you can also take advantage of this option to test a range of values for tool calibration Calibration is important if you are not yet familiar with running Marxan in your study area or you have added new data layers For example by changing the Number of values slider to a number greater than o
25. servation value locking in protected areas can produce solutions with lower total scores than from an un constrained 8 solution Thus when interested is identifying the highest priority parcels without respect to management choosing Available will allow Marxan to select only high value parcels for the final solution regardless of protection status We recommend all users employ this option when comparing existing reserve systems to near optimal designs based on user defined targets see Appendix G for examples 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 prevent Marxan from selecting isolated land parcels but when doing so keep in mind that this has the potential to considerably increase the cost of the final solution Users should also consider the contribution of existing green space to facilitate dispersal in target species because maintaining those
26. slons co etii eti D eee eade t ep tire tes 9 2 2 2 Protection Targets ua e e e ei t Ede Le Pe eo E ELEC AE ERR Ede ee oos 11 2 2 3 Species Penalty Factor et tert e t ET ede haee e PB Rex EET TN nde e gu Hang 12 PAPE tenemur H 13 2 2 5 Number of Iterations trt Rec nee ene a rae e epe e i e Rae ee eA vae ERE cede 13 2 3 Running Marxan and Interpreting the Results ssessssssssseeeee eene nennen nnne enne 14 2 3 Scenario NAMES oue ettet tette ee ether e gettin A decise d akut der destuenterse eens 14 2 3 2 Summary Tabl8Ss irte reet vtt Ath ait ea teet ret tesi A 14 2 373 OUTPUT plots io rete P du DE i MA A eee eas 16 2 4 Downloading and Viewing MARXAN results in ArcMap ccccsessessscecececessesseaececeessesseseaeeeeeesseeeees 17 3 sReferences tlc ted d e UL De doo x Ty A a d Meu RI Meuble o asus e Mar 18 4 Additiorial ResoUrces eret en edet exire ee REX U dh e eee tae e ve ka ke v eaae ehe eee Ee 19 Be Appendices cce sn Rhet I i ene eee eee de eet 20 Appendix A Old Forest Community Occurrence Map cesses ennemi nina nns 20 Appendix B Savannah Community Occurrence Map esses eene nnne enne nnns nnn tinis 21 Appendix C Wetland Community Occurrence Map ccccccccccecsssssssssecececesseseaececeeecessesnaaeseeeesesssessaaees 22 Appendix D Human Commensal Birds Community Occurrence Map
27. t in Arc Map 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 the 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
28. t 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 are 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 17 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 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 Nex
29. 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 2A 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 X 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 nature of your project cost could be calculated as the area of planning units the costs of ongo
30. tof target conservation features species habitats soil types 3 Aclearly defined objective or series of objectives ex 3096 of all grizzly bear habitat 4 How much of each conservation feature is contained within each planning unit The user interface we use in this tutorial 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 BR2CR You will be directed to the following index page Protected Marxan CDFCP
31. which the target s have not been met in the solution for that run expressed as a proportion Missing Values 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 15 2 3 3 Output plots Cost SPF Cost SPF Connectivity Solution Score Iterations Download S3 B1 l1e 05 S3 B10 I1e 05 S3 B100 I1e 05 Cumilative of solutions 110 120 130 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 10096 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 SPF 3 Nreps 100 Niter 1e 05 T 820000 825000 830000 T 805000 810000 81 5000 Boundary length 16 The second tab in the r
Download Pdf Manuals
Related Search
Marxan Tutorial for the Coastal Douglas Fir and