FunConn v1 User`s Manual: ArcGIS tools for Functional Connectivity
Contents
1. FunConn vl Users Guide FunConn v1 User s Manual ArcGIS tools for Functional Connectivity Modeling Authors David M Theobald John B Norman Melissa R Sherburne Contact info Natural Resource Ecology Laboratory Colorado State University Fort Collins CO 80523 www nrel colostate edu projects starmap starmap nrel colostate edu Funding Disclaimer The work reported here was developed under the STAR Research Assistance Agreement CR 829095 awarded by the U S Environmental Protection Agency EPA to Colorado State University This document has not been formally reviewed by EPA EPA does not endorse any products mentioned here Citation Theobald D M J B Norman M R Sherburne 2006 FunConn v1 User s Manual ArcGIS tools for Functional Connectivity Modeling Natural Resource Ecology Lab Colorado State University Fort Collins CO Page 1 7 26 2006 FunConn vl Users Guide Abstract The FunConn pronounced funkin modeling toolbox for ArcGIS v9 provides graph theoretic based analysis methods for landscape connectivity and offers substantial extensions to traditional least cost path approaches Conservation practitioners should embrace this innovative approach for several reasons The graph landscape network data structure allows for an elegant computationally efficient representation of a landscape Complex landscapes can be modeled across large regions encompassing thousands of habitat patches with2. FROMVAL TOVAL V90001 V90002 V90003 V90004 99999999 0 0 0 0 0 0 10 0 0 0 0 10 50 1 1 1 1 50 100 2 2 1 1 100 200 5 6 2 2 200 300 27 8 6 5 300 400 50 22 22 12 400 500 73 50 50 26 500 750 95 73 73 50 750 1000 99 95 95 88 1000 1500 100 100 100 100 1500 2000 100 100 100 100 2000 3000 100 100 100 100 3000 4000 100 100 100 100 4000 5000 100 100 100 100 5000 7500 100 100 100 100 7500 10000 100 100 100 100 10000 999999999 100 100 100 100 Aggregation Factor The input datasets can be aggregated by the given factor for faster processing For example if the aggregation factor is set to 3 with an input resolution of 30x30 the datasets will be aggregated up to 90x90 The output dataset will have the same resolution as the aggregated datasets If you do not wish to aggregate enter a value of 1 but be aware that processing time is inversely related to the aggregation factor and to the square power Different aggregation factors will produce different results For example using a larger aggregation factor will make your data coarser and possibly filter out small patches or generalize patch shape A finer aggregation factor will recognize greater detail It might be useful to experiment with different aggregation factors depending upon the scale of your data and goals of your project You will encounter the aggregation factor twice in the Habitat Modeling process here and in the third tool Build Landscape Network Be awar
3. aa E ines PSRE ESSEEESES gs 43883 558558 N Distance from Patch Edge meters Figure 2 Edge Positive Patch Structure Mountain Lion for example Page 14 7 26 2006 FunConn vl Users Guide Edge Neutral Species Bobcat 100 90 80 707 Distance 60 4 Inside Patch 50 4 40 30 20 4 104 Distance Outside Patch Habitat Quality Patch Edge 1 10 50 50 100 100 200 200 300 300 400 400 500 500 750 750 1000 1000 1500 1500 2000 2000 3000 3000 4000 4000 5000 5000 7500 7500 10000 gt 10000 Distance from Patch Edge meters Figure 3 Edge Neutral Patch Structure Bobcat for example Please note how the habitat guality for the patch structure tables remains at 100 at distances nearby the outside edge of the patch This allows for small stepping stone patches of optimal land cover to be included based on dispersal capabilities of the species Otherwise they would be eliminated from the analysis The distance outside the patch at which habitat quality remains at 100 reflects how far the organism can successfully reach a stepping stone or source patch this should be based on dispersal capabilities Format dbf table Required Fields Case Sensitive 1 FROMVAL short or long integer type 0 decimal places inner distance boundary 1 FROMVAL is patch center 99999 distance units 2 TOVAL short or long integer
4. Any resulting patches from the regiongroup less than the minimum patch size are eliminated amp Il Define Functional Patches Habitat Quality Raster C Mpnxtresultstex_hq gt Minimurn Patch Size ha 264 Patch Foraging Radius meters 917 Core Habitat Percentage 0 1 Resource Quality Threshold 75 Functional Patches Output Raster C Xlynatresultstex patches OK Cancel Environments Show Help gt gt Page 21 7 26 2006 FunConn vl Users Guide Parameters Habitat Quality Raster Enter the Habitat Ouality Raster This raster contains habitat guality data ranging from O to 100 percent acceptable to the target organism and resulted from the preceding tool Create Habitat Quality Minimum Patch Size ha Enter the minimum patch size in hectares the smallest biologically significant patch size for the target organism It may be based on known home range sizes or by estimating home range size using allometric relationships between body mass and home range size Jetz et al 2004 The patch size entered here should be the same as that entered in the Create Habitat Quality Tool If using allometric estimation of home range size we encourage you to run the model atan order of magnitude more and less than this estimation to ensure that the full range of possible home range sizes is covered The minimum patch size for the lynx example based on allometric estimation is 264 ha Patch Foraging Radi
5. Legend ae oe 7 z highways di za a c ade linkages k o z a p E p id ma colors r aes MJ Patches FunConn vl Users Guide PART 2 Landscape Networks Toolset The Landscape Networks toolset enables users to create landscape networks from existing point polygon or polyline datasets The Processing toolbox generates the networks the Analysis toolbox enables the user to calculate a variety of metrics and the Export toolbox allows the user to generate data matrices For Analysis and Export the user can use landscape networks that are generated from their own datasets in the Processing steps or landscape network output from the Habitat Modeling toolset ArcToolbox ArcToolbox Funconn vi Habitat Modeling 3 1 Create Habitat Quality 3 IL Define Functional Patches Z II Build Landscape Network Landscape Networks Sy I Processing II Analysis amp IL Export 3D Analyst Tools Analysis Tools Cartography Tools Conversion Tools Coverage Tools ay Data Management Tools ug Geocoding Tools ag Geostatistical Analyst Tools ag Linear Referencing Tools ug lynx example Spatial Analyst Tools Spatial Statistics Tools Favorites Page 29 7 26 2006 FunConn vl Users Guide I Processing Sub Toolset Points to Landscape Network Tool This tool generates a landscape network based on existing point feature class Attributes of the point
6. you might lose a good amount of the disturbance See Appendix Methods for Creating the Landcover with Disturbance Raster Disturbance Reclass Table Avoidance of roads by organisms especially due to traffic has importance ecological impacts Forman and Alexander 1998 This reclass table attempts to capture these impacts by measuring the effect of a disturbance at given distances The table s values range from 0 total habitat guality loss or 0 guality retained to 100 no effect on habitat guality or 100 guality retained For some initial guidance on guantifying your focal species sensitivity to roads refer to Richard T T Formans text Road Ecology Science and Solutions Format dbf table Reguired Fields Case Sensitive 1 FROMVAL inner distance boundary originates at the disturbance center 99999 distance units e g meters 2 TOVAL outer boundary distance outermost is the furthest possibility 99999 distance units 3 V lt classes gt User defined disturbance class that corresponds with the disturbance class must add a V preceding the class ID i e if Page 17 7 26 2006 FunConn vl Users Guide the class is 90001 in the raster it is V90001 in the reclass table Include a field for each disturbance class Example Table The example table is c lynx disturbance dbf Table 4 Table 4 Example Lynx Disturbance Reclass Table
7. Item s Numeric field s of the point or polygon feature class that will be referenced in the equation string Edge Feature Class This is a valid landscape network polyline feature class use linkages edges or corridors Edge Item s Numeric field s of the polyline feature class that will be referenced in the equation string Calculation Field Name Field associated with the edges linkages feature class that will hold equation string solution This field must be created by the user prior to running the script Equation String DO NOT ADD SPACES ANYWHERE IN EQUATION STRING Network Functions Fields associated with Nodes or Patches must have a leading from field name gt or ato lt field name gt e g from Shape Area to Shape Area ShapeLength Mathematical functions such as pow x y can be entered as pow from_Shape_Area 2 This signifies that the area Shape_Area of the from node is going to be set to the power of two Available Mathematical Functions The syntax of the mathematical functions is as follows x and y should be replaced with feature class field names or constant values O Eo exp x log x log10 x pow x y sin x sqrt x tan x operation separators Example equation strings The average patch area divided by edge linkage corridor length from Shape Area rto Shape Area 2 Shape Length The area of the to Patch multiplied by the Edge length to Shape Area Shape length
8. Modeling and Landscape Networks The Landscape Network toolset contains 3 sub toolsets Processing Analysis and Export Page 7 7 26 2006 FunConn vl Users Guide FEATURE CLASS ATTRIBUTES For Edges Linkages and Corridors Standard attributes such as Object ID FID Shape Area and Perimeter are omitted uk Corridors Linkages Edges Qn Fields q5 Qro G25 Q50 min qmean The value for that line below which qn values fall Shape Length The Euclidean distance from centroid midpoint centroid Straight Dist The Euclidean distance from centroid to centroid without crossing through the midpoint q values C2C_Angle Centroid to centroid angle Mean Angle Vect The average angle across all edge segments Linkages Qn Fields 95 Gro G25 G50 mins mean The value for that line below which qn values fall Shape Length The Euclidean distance from centroid midpoint centroid Straight Dist The Euclidean distance from centroid to centroid without crossing through the midpoint g values B2B Angle Boundary to boundary angle Mean Angle Vect The average angle across all edge segments Corridors Area Area of the corridor calculated in map units Perimeter Length of the entire corridor boundary Topo num Topological number is the number of patches that connect via that corridor One corridor can connect many patches Page 8 7 26 2006 FunConn vl Users Guide From SharBound The w
9. along the x axis from dso to the point at which y 0 75 McGarigal 2001 Permeability 1 0 1 0 1 0 e sope 450 ds dso 30 d 7 1 0 8 E 0 6 iy E 0 4 a 0 2 0 0 10 20 30 40 50 60 70 80 90 Slope degrees Figure 4 Permeability logistic function As with the permeability reclass table the range of values is 0 1 with 1 being the most readily moved through The slope cost is multiplied by the permeability raster to get a total permeability raster The final resistance or cost surface is simply the inverse of the permeability raster Low cost example A low cost example for lynx is a cell that is spruce fir and has a slope of 20 Spruce fir has a permeability of 1 0 20 has a permeability of 0 81 so that cell s total permeability is 1 0 x 0 81 0 8 The permeability is inverted to obtain the cost resistance surface 1 0 0 8 1 23 High cost example Using the same 20 slope but a more impermeable land cover type such as short grass prairie permeability lt 0 20 greatly increases the total cost for that celi Page 26 7 26 2006 FunConn vl Users Guide Total permeability 0 81 x 0 2 0 162 Cost weight resistance 1 0 0 162 6 17 Aggregation Factor ENTER THE SAME AGGREGATION FACTOR AS USED IN THE CREATE HABITAT GUALITY TOOL See Create Habitat Ouality Tool for more information qn Value Specify the n value the cost allocation value
10. habitat_quality dbf tri Resource Ouality Threshold 75 Minimum Patch Size ha 264 Patch Structure Reclass Table Ctlynmipatch structure dbf Land Use Cover with Disturbance Raster C Xlpnakvp disturb Disturbance Reclass Table CAlynekdisturbance dbf pe Aggregation Factor 3 Habitat Ouality Output Raster Ce Xipnasresultsklynx hg OK Cancel Environments Show Help gt gt Page 11 7 26 2006 FunConn vl Users Guide Parameters Land Cover Raster Specify the land cover raster that represents the landscape to be used by your organism This can be USGS NLCD USGS GAP or user defined but must be in raster format and be categorical or class nominal data The example raster is c lynx vp_swrgp and is derived from the Southwest Regional GAP data http fws nmcfwru nmsu edu swregap There is a 4000 character limit within ArcGIS software for the raster conditional statement used within this tool f the file or path name is too long or if your raster has a large number of classes such as Southwest ReGap with 125 this may cause this tool or process to fail because the conditional statement exceeds the 4000 character limit To avoid this problem save your land cover and disturbance rasters directly in c or c temp with short file names and or reduce the number of classes in your land cover dataset Resource Quality Reclass Table This table contains the data necessary to reclassif
11. the shortestpaths table will contain only the paths that are the shortest between the source node Page 40 7 26 2006 FunConn vl Users Guide and all other nodes If you choose to increase this to say 1 1 then all paths from a source node to each other node that are the length of the shortest path plus 10 of that distance are included in the table For example if the shortest path from node A to node D was ABD and equal to a distance of 10 that is the only path between A and D the table would return using a default threshold factor of 1 However if a threshold factor of 1 5 is used than any path with a distance egual to or less than 15 would also be included in the table So then if path ACD 13 and path ABCD 15 they would also both be included in the table Therefore for the node pair AD the table will contain 3 paths instead of just the shortest Be aware that increasing the Threshold Path Weight Factor will drastically increase your computing time Add Path List Field optional This optional field allows from the node or patch FID s that fall along a given path to be recorded in the field Path List If the path exceeds 255 characters the full linkage path will not be fully recorded Page 41 7 26 2006 FunConn vl Users Guide lll Export Sub Toolset Node Edge Node Distance Matrix D This tool produces a matrix of distances dj between 1 order neighbors nodes and You can use edg
12. the user specified q value a certain percentage of the cell groups are extracted and serve as mid points for a set of initial linkages between patches The initial linkages are joined from each q cell group to the nearest point on the source patch A second set of allocation zones are grown outward from the initial linkages where they meet are considered fences and are removed from each allocation zone so that unique corridors cannot merge between patches Within each new allocation zone the cells less than the specified q value are pulled out to form the corridors The least cost cell at the mid point and patch boundary of each corridor is found to create the refined linkages Multiple edges are formed between each pair by using the same mid point as the linkage The edges are then formed by straight line segments from patch centroid to midpoint to patch centroid 5 III Build Landscape Network Functional Patches Raster Land Use Cover with Disturbance Raster Landcover Permeability Reclass Table Additional Permeability Raster optional Aggregation Factor Links Qn Value 10 Output Landscape Network Cancel Environments Show Help gt gt Page 24 7 26 2006 FunConn vl Users Guide Parameters Functional Patches Raster Patches can either be the output from the preceding tool Define Functional Patches or an integer raster with unigue patches generated by t
13. type 0 decimal places distance to the next boundary last TOVAL is outermost possibility 99999 distance units 3 QUALITY short or long integer type O decimal places habitat quality at a distance class value range 0 not habitat to 100 optimal habitat Example Table c lynx patch_structure dbf Table 2 Page 15 7 26 2006 FunConn vl Users Guide Table 2 Example Lynx Patch Structure Reclass Table FROMVAL TOVAL QUALITY 99999999 2000 100 2000 1000 100 1000 500 100 500 400 100 400 300 100 300 200 90 200 100 70 100 50 50 50 0 25 0 10 100 10 50 100 50 100 100 100 200 100 200 300 100 300 400 100 400 500 100 500 750 90 750 1000 75 1000 1500 50 1500 2000 40 2000 3000 25 3000 4000 10 4000 5000 5 5000 7500 0 7500 10000 0 10000 999999999 0 Land Cover with Disturbance Raster This raster incorporates disturbances such as roads agricultural areas and high intensity development into land cover data In many land cover datasets these disturbances are grouped into a general development class ignoring the varying magnitude of different types The example raster is cAlynxp disturb Note This is nota raster of only the disturbances it is a raster of land cover with disturbances included In many land cover datasets roads or high intensity development are grouped into one class suc
14. used to generate the selection set Weight Field optional This is a numeric field for edge feature class that will be used as the distance weight in then selection process This field is only necessary if the neighborhood metric is set to Edge Distance Weight Threshold Value Specify the value that will limit the analysis Page 36 7 26 2006 FunConn v1 Users Guide Node Calculator Tool This tool evaluates a user defined eguation string that is executed for every node in a Landscape Network feature class populating a defined field with the solution The equation string can contain elements fields from first to n order nodes and edges for a given source node This is possible through syntax that distinguishes source nodes i from first order nodes j and edges e and n order neighborhood and paths summation statements and other common mathematical functions exp x The equation string must reference fields selected for a given feature class with a prefix of i source node j first order nodes e first order edges k n order neighborhoods and p shortest paths for a given weight threshold If first order node edge neighborhood or path fields are referenced they must be surrounded within the summation syntax brackets because there is usually more that one first order node or edge resulting in the need to sum their values The summation syntax can be nested and incorporate source node i fiel
15. 26 2006 Example Lynx Habitat O 20 Miles uality Vail Pass Colorado A A oes s highways Habitat Quality High 100 Low 0 FunConn v1 Users Guide ll Define Functional Patches This tool defines functional patches based on the focal species minimum foraging requirements and ability to move between patches The habitat quality raster and four user defined parameters maximum foraging radius minimum patch size core habitat percentage and resource quality threshold guide this process Methods Overview All areas greater than the habitat quality threshold default 75 are kept and regiongrouped according to an eight neighbor rule in GIS terms this means that adjacent cells including diagonal directions are grouped together If these areas are less than the core habitat percentage times the area of the foraging radius for the example lynx this is 0 10 1 917 they are eliminated Regions or patches that are larger are kept and called the core seed areas A cost surface is created from the habitat quality raster Cells of high quality habitat have very low cost cells of poor quality have high cost From the seeds patches are grown outward across the cost surface to a cost distance equal to the foraging radius e g 917 meters for example lynx By growing the patches away from the seeds across a cost surface the spatial variation and complexity is retained
16. Least cost distance between nodes Cij pow qncost 1 75 Page 35 7 26 2006 FunConn vl Users Guide Neighborhood Selection Tool This tool generates a selection set of nodes or edges based a user defined neighborhood or edge weight threshold value The selection analysis originates or grows from an existing selection set of nodes If there is not an existing selection set all nodes in the network will be evaluated To view the new selection set the ArcGIS viewing window has to be refreshed manually after the tool has completed Neighborhood Selection Node Feature Class Edge Feature Class Features to be Selected Neighborhood Metric Weight Field optional Threshold Value Cancel Environments Show Help gt gt Parameters Node Feature Class This is a landscape network point or polygon feature class This could be nodes or patches Edge Feature Class This is a valid landscape network polyline feature class This could be linkages edges or corridors Features to be Selected A Landscape Network node patch edge or linkage feature class that selected features will be selected from Neighborhood Metric A neighborhood metric that determines how the selection set will be processed If Neighborhood Order is selected then topological relationships between nodes will be used to in the analysis If Edge Distance Weight is selected then an edge or linkage field will be
17. The Processing toolset generates the landscape network based on points polygons or polylines The Analysis toolset allows for graph theoretic or network type analyses to be executed on landscape networks The Habitat Modeling landscape network is suitable for analysis within this toolset Tools included in the Analysis toolset allow for calculating minimum spanning trees based on a user defined weight values calculating node and edge interactions based on user defined fields and equation strings and finding the shortest paths from each node to every other node in the network The Exporttoolset exports the landscape network to an NxN matrix based on user defined weight values An example dataset is provided so you can become familiar with the FunConn concepts and tools before applying them to your datasets These data are in the ynx folder save this to your local drive Page 2 7 26 2006 FunConn vl Users Guide TABLE OF CONTENTS Software Environment Example Dataset Terminology Feature Class Attributes PART 1 HABITAT MODELING TOOLSET I Create Habitat Quality Il Define Functional Patches lll Build Landscape Network PART 2 LANDSCAPE NETWORKS TOOLSET l Processing Points to Landscape Network Polygons to Landscape Network Polyline to Landscape Network Il Analysis Minimum Spanning Tree Edge Calculator Neighborhood Selection Node Calculator Shortest Paths Ill Export Node Edge Node Distance Matrix D Node Pat
18. X Methods for Creating the Landcover with Disturbance Raster To create a landcover raster with disturbance you will need two datasets 1 Roads attributed feature class 2 Landcover raster STEP NI STEP 2 Assuming that your roads will be a feature class there are 3 steps to creating the new raster STEP 1 Convert roads to a raster based on some characteristic found in the attributes e g road type or traffic volume Feature to Raster Input features la HIGHWAYS Field SAD T v Output raster C temp hwy_rast Output cell size optional so Cancel Apply Show Help gt gt STEP 2 Reclassify the road raster into 4 5 classes These classes must be different from existing land cover classes In this example we use 90001 90005 Page 45 7 26 2006 FunConn vl Users Guide Reclassify Input raster les Hw Raster Reclass field VALUE Reclassification raje z Old values New values A 0 8553 515625 90001 Classify 8553 515625 20412 5 90002 on 30412 5 67477 734375 90003 Unique 67477 734375 136856 25 90004 136856 25 243300 90005 Add Entry NoData NoData Delete Entries vi Load Save Reverse New Values Precision Output raster C temp aadt_reclas Change missing values to NoData optional v Cancel Apply Show Help gt gt Classification Classification m Classific
19. _Area Fields associated with first order neighboring node of source node must be prefixed with j e g j Shape_Area Fields associated with first order edges must be prefixed with e and a grouping command e g e Shape_Length avg The character vin the equation represents the number of first order nodes e g j Shape_Area i Shape_Area k 1 The character k is used to access node or edge attributes associated with a given topological neighborhood order e g k Shape_Area 3 or k Shape_Area 3 Shape_Length The example equation string k Shape_Area 3 finds all nodes within three neighborhood orders of node i and sums their Shape_Area attribute Note that the character n can be used to calculate network wide metrics the n option does not include the source node i e g k Shape_Area n i Shape_Area The character p is used to access node or edge attributes associated with a given path weighted neighborhood e g p Shape_Length 5000 or p Shape_Length 5000 Shape_Area The example equation string p Shape_Length 5000 finds all nodes within 5000 Shape_Length units of node i and sums the Shape_Length path values Note that the character n specified to calculate network wide metrics from source node i e g p Shape_Length n The character m is used to get the number of nodes associated with a given topological neighborhood order of node i e g m 3 The example equation string m 3 will return the
20. ach unique land cover class Example Table The example table is c lynx permeability dbf Table 5 Table 5 Example lynx permeability classes from cAlynxwp perm VALUE DEFINITION PERMVALUE 22 Rocky Mountain Aspen Forest and Woodland 1 000000 32 Rocky Mountain Montane Mesic Mixed Conifer Forest and Woodland 1 000000 41 Rocky Mountain Gambel Oak Mixed Montane Shrubland 0 800000 68 Chihuahuan Gypsophilous Grassland and Steppe 0 200000 83 North American Warm Desert Riparian Woodland and Shrubland 0 100000 7 Western Great Plains Cliff and Outcrop 0 100000 17 North American Warm Desert Active and Stabilized Dune 0 100000 28 Rocky Mountain Subalpine Mesic Spruce Fir Forest and Woodland 1 000000 90005 Developed Medium High Intensity Class 112 in the swrgp_funconn 0 005000 Page 25 7 26 2006 FunConn vl Users Guide Additional Permeability Raster The permeability raster should represent a surface that would change the rate of movement between functional patches such as slope terrain ruggedness or some other topographic index The example permeability surface which is derived from slope is cilynxivp slope perm A negative logistic function was used to create the example slope cost raster Two parameters determine the shape of the curve the inflection point dso and the scaling factor ds The inflection point is the value of x where y 0 5 The scaling factor is the distance
21. and then populates a defined field with the solution The eguation string can contain elements fields from nodes patches and edges linkages for a given source edge This is possible through syntax that distinguishes node patch and edge linkage attributes To distinguish node patch attributes from edge linkage attributes node patch attributes have a leading from lt field name gt or ato lt field name gt where as the edge linkage attributes are just the field name The leading from and to_ are to designate from and to node patch attribute values within the equation string The equation string can also incorporate other common mathematical functions and constant values The solution of the equation string is storied in a pre defined field associated with an edge of linkage feature class The user must create this empty field within the edge linkage feature class before executing the equation Examples include finding node weight interaction and weighting edges by node attributes amp Edge Calculator Node Weight Feature Class Node Weight Item s Le le t lx le Edge Feature Class Edge Item s KU le la lx le Calculation Field Name Y Equation String v Cancel Environments Show Help gt gt Page 34 7 26 2006 FunConn vl Users Guide Parameters Node Feature Class Select a landscape network point or polygon feature class This could be nodes or patches Node Weight
22. atch Path A walk in which all nodes and edges are unique If a path has more than 3 nodes with no cycles it is a tree Walk A sequence of nodes connected by edges If a walk ends at the first node it is a cycle Core Seed An area of high guality habitat from which the functional patches originate From the seed the patches are grown across a cost surface to a distance egual to the units of the foraging radius See Define Functional Patches tool Model The definition of a model depends on its context 1 Simulation of a process or response at a given scale For example lynx natal dispersal across a landscape i e lynx habitat model v modeling 2 In ArcGIS geoprocessing a process or series of linked processes represented by a flow diagram in Modelbuilder Tool A tool is a Python script found in a toolset The Habitat Modeling toolset contains 3 tools Create Habitat Quality Define Functional Patches and Build Landscape Network The Landscape Network Processing toolset contains 3 Page 6 7 26 2006 FunConn vl Users Guide tools Points to Landscape Network Polygon to Landscape Network and Polyline to Landscape Network The Landscape Network Analysis toolset contains 5 tools Calculate Minimum Spanning Tree Edge Calculator Node Calculator Neighborhood Selection and Shortest Paths Toolbox The entire collection of FunConn toolsets and tools Toolset FunConn contains 2 primary toolsets Habitat
23. ation Statistics Method Natural Breaks Jenks y Count 1273730 y Minimum 0 Classes JE z Maximum 243300 Sum 4479698524 21875 Mean 3516 992239 Data Exclusion Exclusion Sampling Standard Deviation 12193 787148 Columns 100 JO Show Std Dev Show Mean 7 Break Values xl 8553 515625 30412 5 67477 734375 136856 25 243300 1500000 67477 7343 1000000 500000 60825 121650 182475 243300 Snap breaks to data values Cancel Page 46 7 26 2006 FunConn vl Users Guide STEP 3 Use a conditional statement to embed the roads into the landcover raster Single Output Map Algebra Conditional Map Algebra expression conlisnul AADT Reclass SWREGAP 44DT_Reclass Output raster jc temp swrap_disturb k Input raster or feature data to show in ModelBuilder optional Input raster or feature data optional zl Es HWY Raster Reclassed by AADT Es SWREGAP le Lx k v Cancel Apply Show Help gt gt Page 47 7 26 2006 Habitat Modeling Geoprocessing Steps Overview The Geoprocessing processes behind the three primary steps to the FunConn Habitat Modeling process are outlined here The three primary steps are 1 The Create Habitat Ouality step creates an initial habitat guality raster 2 The Define Functional Patches step produces a rasters of discrete patches 3 The Build Landscape Network step cr
24. ches the edge In the case of lynx an old growth spruce fir patch is optimal however the core of the patch is more valuable than the edge These core favoring species are edge negative Figure 1 Other species such as mountain lion might prefer edge areas due to their hunting practices The patch structure table would reflect that by having slightly lower habitat quality values for the inner most area of the patch These species are edge positive Figure 2 Species that exhibit no preference for core or edge habitat areas are edge neutral Figure 3 Page 13 7 26 2006 FunConn vl Users Guide Edge Negative Species Lynx sl 90 4 el 80 s al 70 4 Distance l Distance S 604 Inside Patch Outside Patch o 507 I o 407 304 20 104 0 TO a a S28 28S 28 o ZE es sa ma 2 2 e T e D o i i 1 i re TON o v bb m o O Sr Ge re A ae io O DO S SM SD E ce E E JA A LS SS Ss 2aossS 8 8 8 8 8 88938 OS ai IC Fees 888 8 O Mm Loro A ns y Distance from Patch Edge meters Figure 1 Edge Negative Patch Structure Lynx for example Edge Positive Species Mountain Lion 100 a 90 I 80 l l z 707 T 604 Distance Distance Inside Patch Outside Patch 507 l el o 401 Si ul T 30 si 20 al l 10 4 I See ese Ls ee SVE eee Se al Y e lca o o o i 1 a a a i se N ma vt tej m o o dale O A A See o o o o o o A j E E E
25. count of third order neighbors around node i The character n is used to get the total number of nodes in the network minus the source node i e g k Shape_Area n i Shape_Area n 1 brackets are used to indicate that the function inside is to be summed up before proceeding e g i Shape_Area Page 38 7 26 2006 FunConn vl Users Guide Enter mathematical functions such as pow x y as pow i area 2 This signifies that the area of source node is going to be sguared Available Mathematical Functions exp x log x log10 x pow x y sin x sgrt x tan x operation separators Example eguation strings Get the average first order node area for a given source node j Shape Areaj v Calculate the natural log of first order nodes times the edge that connects them to the source node length divided by the source node area sguared log 4j Shape Area l Shape Length pow i Shape Area 2 Calculate the average node Shape Area for fourth order nodes around node i k Shape_Area 4 m 4 Calculate the sum of Shape Length paths with a threshold distance of 5000 Shape Length units p Shape Length 5000 Calculate the average Shape Length paths from source node i to all other nodes within the network 4p Shape Length nj n Page 39 7 26 2006 FunConn vl Users Guide Shortest Paths The Shortest Paths tool finds the shortest distance from a source node to all other nodes in the
26. ds and other common mathematical functions The syntax for the mathematical functions listed with the equation string help entails using the supplied function string replacing the x or y values with fields or constant values e g pow i Shape Area 2 Node Calculator Node Weight Feature class Node Weight Item s E le lx le Edge Feature Class Edge Item s bd E le lo lx Calculation Field Name Equation String Cancel Environments Show Help gt gt Page 37 7 26 2006 FunConn vl Users Guide Parameters Node Weight Feature Class This is a landscape network point or polygon feature class This could be nodes or patches Node Weight Item s Specify the numeric field s that will be referenced in the equation string Edge Feature Class This is a valid landscape network polyline feature class This could be edges linkages or corridors Edge ltem s Numeric field s of the polyline feature class that the equation string references Calculation Field Name Field associated with the edges linkages or corridors feature class that will hold equation string solution The user must define this field prior to running the script Equation String DO NOT ADD SPACES ANYWHERE IN EQUATION STRING Network Functions Elements within equation must be defined with s You must prefix the fields associated with source node in the equation with i e g i Shape
27. e that once you aggregate up you cannot aggregate down to a smaller cell size you must be consistent both times For example if you aggregate your land cover raster from 30 meter resolution to 90 meter resolution in the Page 18 7 26 2006 FunConn vl Users Guide Create Habitat Quality Tool you cannot decide to change this to 60 meter resolution in the Build Landscape Network Tool If you foresee wanting to have an multiple landscape networks at different scales the best approach would be to generate your Habitat Quality and Patches rasters at the finest grain e g 30 meters and use different aggregations in the Build Landscape Network Tool It is beneficial to test different aggregation factors on a sample area of your study area before deciding on the best one Habitat Quality Output Raster Specify the name and location of the output raster This raster is the input for the next tool Define Functional Patches The example output is cilynxiResulisihabitat quality The habitat quality directory is created using hg following with a time stamp cXtempihg lt yyymmddhhmmss gt This time stamp directory contains a variety of temporary rasters although the final habitat guality raster is saved in the location that you specify See the readme file in the hg directory folder for a description of its contents Within the time stamp folder is a README file with the parameter specifications of the model run Page 19 7
28. eates functional connections between patches which are represented by linkages and corridors LEGEND z raster dataset input or output geoprocessing function input parameter value integer input parameter table user defined dbf geodatabase Note These diagrams capture the general flow of processes but do not include every geoprocessing step especially the case in the Build Landscape Network Tool diagram For the exact steps refer to the Python script Habitat Modeling Geoprocessing Steps 1 Create Habitat Quality Tool m i a Conditional Reclassification Reclassed ggrega ion to Agr td Reclassed Statement TI Landcover a larger cell size Landcover Regiongroup Habitat Quality 0 100 All land cover Positive Initial Negative groups gt 75 Conditional Patch Conditional Patch Euclidean gt min patch size Statement Raster Statement Raster Distance Distance inside Distance outside patches patches Conditional Statement Distance inside and outside patches e Map Algebra mi m Multiplied by Resource If gt 100 ULA Reclassification quality and disturbace setto 100 Dvided by 10000 RASTER Patch Structure Habitat Modeling Geoprocessing Steps 2 Define Functional Patches Tool Core HABITAT Xi vata ra Z guality GUALITY Conditional vata ra Z areas Conditional Habitat Seeds RASTER Statement gt 75 Statement gt seed area ED habitat lt oraging radiu X threshold Core seed area d
29. ef ED SUL map Algebra FUNCTIONAL inverted Ie Cost Distance PATCHES RASTER exp function lt oraging radiu Habitat Modeling Geoprocessing Steps 3 Build Landscape Network Tool FUNCTIONAL Conversion to PATCHES shapefile Patches shp Land Cover 7 w Disturbance Reclassification Permeability X Slope cost aa Cost Raster Permeability Reclass Allocation Zones Alloc Boulndaries Regiongrouped Palais a ere Allocation Zones grown outward Derived where by adjacent a arest ga ge Initial Linkages grown outward cells of 2 intersection point until they meet until they meet zones meet Rana Fences removed from A pene CORRIDORS CORRIDORS the allocation zones E zones meet allocation zone REFINED Landscape east cost cell found at midpoint of each corridor LINKAGES 4 Network mdb and at each edge intersection
30. es linkages or corridors to calculate distances Node edge node distance matrix D Node Featureclass Edge Link Featurecalss Distance Field Output ASCII Matrix CSV Cancel Environments Show Help gt gt Parameters Node Feature Class This is a valid landscape network node feature class This could be nodes or patches Edge Link Feature Class This is a valid landscape network edge featureclass This could be linkages corridors or edges Distance Field Any numeric field representing a distance The distance between each pair of adjacent nodes is calculated distances will be symmetric unless directionality is incorporated Output File Name Specify the location of your output matrix Page 42 7 26 2006 FunConn vl Users Guide Node Path Node Distance Matrix D This tool produces a matrix node to node distances by paths of several edges or linkages This allows for distances to n order neighbors to be included The matrix is exported in a comma delimited format Node path node distance matrix D Node Featureclass Edge Link Featureclass Distance Field Output ASCII Matrix CSV Cancel Environments Show Help gt gt Parameters Node Feature Class This is a valid landscape network node feature class This could be nodes or patches Edge Linkage Feature Class This is a valid landscape network line feature class This could be
31. g Geostatistical Analyst Tools BF Linear Referencing Tools ag lynx example ag Spatial Analyst Tools ag Spatial Statistics Tools Favorites Page 10 7 26 2006 FunConn vl Users Guide I Create Habitat Quality The Create Habitat Quality Tool creates a habitat quality raster that is based on the guality of foraging resources with values ranging from 0 not habitat to 100 optimal habitat Habitat guality is based on three factors resource guality patch structure and distance from disturbance Resource quality reflects the in situ vegetation land cover and is based on species vegetation affinities Patch structure accounts for the so called edge effect by evaluating proximity to patch edge to define core habitat areas Distance from disturbance guantifies the effects of common land use disturbances such as roads and development on habitat guality Methods Overview The land cover raster is reclassified into resource quality values by the resource guality reclass table The surface of the study area is reclassified according to distance from disturbance disturbance reclass table The surface of the study area is reclassified again according to patch structure patch structure reclass table The three resulting rasters are combined to create the habitat quality raster Create Habitat Quality Land Use Cover Raster jc Mynxvp_swrap 4 Resource Quality Reclass Table C lynx
32. h Node Distance Matrix D References Appendix Methods for Creating a Disturbance Raster Geoprocessing Steps Overview Diagram Page 3 7 26 2006 o cd o w 11 21 24 29 30 30 31 32 33 33 34 36 37 40 42 42 43 44 45 45 48 FunConn vl Users Guide Software Environment FunConn tools were written within Python v2 1 as a Geoprocessing toolbox and ArcGIS v9 1 An ArcINFO level license is required to run certain FunConn tools in which a personal geodatabase is created which are the Habitat Modeling Build Landscape Network Tool and the Landscape Networks Processing Tools Once a landscape network is created all other tools can run using either an ArcView or ArcINFO license The FunConn tools also reguire the Spatial Analyst extension Before using the FunConn tools be sure to have the Spatial Analyst extension activated and a c temp directory on your machine this is where output directories are stored All input data should be in the same projected coordinate system and datum To use the Habitat Modeling tools you must have a land cover dataset for your study area All of the additional parameters are user defined and explained in this guide Example Dataset Description The following datasets are found in the lynx folder provided with the FunConn Tools Save this folder to your local drive exactly as c llynx to avoid model failure due to ESRI character limits Use this dataset to run example
33. h as unvegetated To account for disturbance effects these must be categorized more specifically When devising your disturbance classes make them higher numbers than your other vegetation classes This is because when the tool aggregates the cell will default to the highest value If the disturbance classes are low values and you aggregate you might lose a good amount of the disturbance Page 16 7 26 2006 FunConn vl Users Guide One possible technigue for creating this dataset is to embed the disturbances into the land cover dataset To do this first classify the roads into 5 classes based on road type highway local road etc or by traffic volume Average Annual Daily Traffic AADT Next convert the roads into a raster by road class Lastly use a conditional statement to replace the original land cover values with the disturbance grid values The example dataset classifies disturbances as classes 90001 90005 by AADT and high intensity development Table 3 Table 3 Example Lynx Disturbance Classification Assigned Value 90001 0 SK AADT 90002 5K 10K AADT 90003 10K 30K AADT 90004 gt 30K AADT Class 90005 High intensity development When devising your disturbance classes make them higher numbers than your other vegetation classes This is because when the tool aggregates the cell will default to the highest value If the disturbance classes are low values and you aggregate
34. he user Land Cover with Disturbance Raster This must be a categorical raster dataset that contains natural land cover and features that represent disturbances to the target organism such as roads urban areas agricultural areas etc This raster can be USGS NLCD USGS GAP or user defined with disturbances such as roads e g by road type or traffic volume burned into it The example raster is cAlynxivp disturb The raster is the same as used in the Create Habitat Quality Tool Land Cover Permeability Reclass Table This table reclassifies the land cover with disturbance raster based on how easily an organism can move through a given class This is not the same as habitat quality For example a land cover class might have a very low habitat guality value but still is highly permeable In the case of lynx ponderosa pine woodland does not offer valuable foraging or denning habitat but does provide excellent visibility and escape cover so is readily moved through Keep in mind that permeability is the inverse of resistance Permeability is multiplied by the additional permeability raster optional and then inverted to generate the cost weight surface of resistance values Format dbf table Required Fields Case Sensitive 1 VALUE short or long integer O decimal places unique class values for each land cover type 2 PERMVALUE short or long integer 6 decimal places ranging from 0 000000 1 000000 permeability value for e
35. idth of the corridor where it intersects the from patch boundary To SharBound The width of the corridor where it intersects the to patch boundary Mid SharBound The width of the corridor at the center most point of the corridor Page 9 7 26 2006 FunConn vl Users Guide PART 1 Habitat Modeling Toolset The Habitat Modeling Toolset allows you to create a network of functionally defined resource patches for a focal species The toolset consists of three successive tools Create Habitat Guality Define Functional Patches and Build Landscape Network The primary input for these is a land cover raster dataset and a series of parameter tables The parameter tables are user defined so it is important to follow the formatting guidelines explained in this guide and illustrated in the example lynx dataset The output from the final step Build Landscape Network is a geodatabase containing feature classes for nodes patches edges linkages corridors and relationship tables Once you generate the landscape network you can analyze its overall connectivity using the Landscape Network Analysis Tools ArcToolbox ArcToolbox ay FunConn v1 amp 3 1 Create Habitat Quality Z II Define Functional Patches Z III Build Landscape Network Landscape Networks db 3D Analyst Tools ay Analysis Tools ig Cartography Tools ay Conversion Tools EY Coverage Tools ag Data Management Tools Geocoding Tools u
36. is to create a polyline shapefile in ArcMap Do this by creating an empty shapefile in ArcCatalog bringing it into ArcMap and using the editing tools to form it into the desired shape Be sure to have your snapping parameters set to snap to ends Then simply draw line segments The endpoints of the lines will be nodes once you run the Polyline to Landscape Network Tool Digitizing direction is not important because the tool duplicates each edge in the relationship table Please note that ArcGIS can place locks on files that are difficult to detect and will prevent proper execution of FunConn If you experience this problem try restarting ArcMap and refreshing your data 5 Polyline to Landscape Network o Polyline o Output Landscape Network E Cancel Environments Show Help gt gt Parameters Polyline Feature Class Enter a polyline feature class that will be converted into a landscape network Output Landscape Network Enter the name and location of the output landscape network Page 32 7 26 2006 FunConn v1 Users Guide ll Analysis Sub Toolset With all Analysis Tools it is important to know that linkages and corridors can be used as edges and patches can be used as nodes Minimum Spanning Tree Tool In graph literature a spanning tree is a tree more than 3 nodes no cycles that includes every node in the graph Since there may be several spanning trees per graph the minimum spanning tree is the o
37. linkages corridors or edges Distance Field Any numeric field representing a distance The tool sums the distances between adjacent nodes to obtain the total path distance Distances will be symmetric unless directionality is incorporated Output ASCII Matrix Specify output location for matrix Page 43 7 26 2006 FunConn vl Users Guide References Forman R T T 2003 Road Ecology Science and Solutions Island Press Forman R T T and L E Alexander 1998 Roads and their major ecological effects Annual Review of Ecology and Systematics 29 207 231 Harary F 1969 Graph theory Addison Wesley Reading Massachusetts USA Jetz W C Carbone J Fulford and J H Brown 2004 The scaling of animal space use Science 306 266 268 Kruskal J B 1956 On the shortest spanning subtree and the traveling salesman problem Proceedings of the American Mathematical Society 7 48 50 McGarigal K 2001 CAPS documentation Biodiversity Assessment Summary http www umass edu landeco research caps caps html Theobald D M 2006 Exploring the functional connectivity of landscapes using landscape networks In Conservation connectivity Maintaining connections for nature K R Crooks and M A Sanjayan eds Cambridge University Press Urban D and T Keitt 2001 Landscape connectivity A graph theoretic perspective Ecology 82 5 1205 121 Page 44 7 26 2006 FunConn vl Users Guide APPENDI
38. models or to compare with your results Name Format Definition 222222 I pas shapefile disturbance permeability permeability reclass table vail pass counties snape e Vail_pass_highways shapefile highways ESRI s character limit for conditional statements is 4000 Long path names or a large number of classes can cause this to be exceeded Non essential for running the FunConn tools but are provided for reference purposes only These are examples of successful tool completion using the default parameters and example tables use for comparison against the results of your operations Page 4 7 26 2006 FunConn vl Users Guide Terminology Graph A graph is a data structure comprised of a set of points nodes functionally joined by lines edges The set of nodes is typically defined as V G V1 V2 V3 Vp and the edges as E G e1 e gt es eg Therefore the graph G has p nodes order and q edges value G p g Edge e connects adjacent nodes v and v In this application nodes represent habitat patches and edges represent lines of movement For classic literature on graph theory refer to Harary s 1969 text Graph Theory Otherwise Urban and Keitt 2001 provide a comprehensive overview of the application of graph theory to landscape ecology A basic understanding of graph terminology is helpful before using FunConn Planar Graph Nodes edges Non Planar Graph Nodes edge
39. nctional Patches Raster This raster is the input for tool proceeding tool Build Landscape Network Page 22 7 26 2006 Example Lynx Functional Patches Vail Pass Colorado Legend McCoy highways patches unigue colors dark areas in the center of the patches are high guality core i E Jefidrs n i va z areas small black areas outside of the patches are high quality but not large enough to be viable patches FunConn v1 Users Guide HI Build Landscape Network The Build Landscape Network Tool creates a landscape network representing habitat patch connectivity This is accomplished by using functional habitat patches as source regions and land cover data as a movement resistance surface The landscape network is comprised of nodes patches edges linkages corridors and relationship tables Methods Overview The land cover with disturbance raster is reclassified according to user defined permeability values found in the permeability reclass table This raster is multiplied by an optional permeability raster e g slope perm and then inverted to create a cost surface for the study area Allocation zones are grown away from the source patches across the cost surface until they meet The meeting point of the allocation zones are called allocation boundaries These boundaries are 2 cells wide and are comprised of a distribution of cost distance values Depending on
40. ne with the shortest distance By applying ecological weights to edges the minimum spanning tree can represent the graph s parsimonious backbone for conservation Urban and Keitt 2001 In other words the minimum spanning tree is the most efficient route for an organism to disperse through all habitat patches of a landscape This tool creates a table in a landscape network that contains all edges that make up a minimum spanning tree using a greedy algorithm Kruskal 1956 The Kruskal algorithm is executed by adding n 1 least cost edges in ascending order without creating any circuits The parameters for this tool are an edge feature class and a numeric field that the minimum spanning tree will be calculated from The result is a selected set of edges you must manually refresh the screen to view results Calculate Minimum Spanning Tree o Polyline Featureclass o Polyline Weight Field OK Cancel Environments Show Help gt gt Parameters Polyline Feature Class This must be a valid landscape network polyline feature class either edges linkages or corridors Polyline Weight Field This must be a numeric field for the polyline feature class The numeric field could be length gncost or any other weight attribute Page 33 7 26 2006 FunConn v1 Users Guide Edge Calculator Tool This tool evaluates a user defined eguation string that is executed for every edge in a Landscape Network feature class
41. network The tool produces a shortestpaths table with four fields sourcenode The sourcenode field is the starting node or patch FID and tonode the tonode is the ending node or patch FID of a given path pathlist The pathlist field holds the sequence of nodes starting with the sourcenode moving to the tonode This field is optional and only shows linkages that are less than 255 characters long sumweights The sumweights field is the total distance from the sourcenode to the tonode Shortest Paths Node Feature Class Edge Feature Class tae Path Weight Field Y Output Table Name must be located in Landscape Network Threshold Path Weight Factor add Path List Field optional v Cancel Environments Show Help gt gt Parameters Node Feature Class This is a valid landscape network point or polygon feature class This could be nodes or patches Edge Feature Class This is a valid landscape network polyline feature class This could be linkages corridors or edges Path Weight Field A numeric field found within an edge or linkage feature class that is used to calculate the paths between nodes or patches Examples include any distance length or cost attribute Output Table Name Specify the output location must be located within the existing landscape network Threshold Path Weight Field This is a numeric value greater than or equal to one The default factor is 1 By leaving it as 1
42. out compromising fine grain spatial variation Users can evaluate landscape level connectivity between all patches not simply adjacent patches 1 order neighbors This is a valuable way to identify bottlenecks locations that are critical for overall connectivity due to the spatial configuration of habitat not simply the influence of immediate neighborhood context Graph edges landscape network linkages and corridors can have user defined weight attributes providing flexibility in calculating a variety of graph based landscape metrics such as patch area corridor width and average cost distance In short the graph theoretic approach to modeling landscape connectivity is more flexible efficient and more powerful than traditional least cost path analysis This document provides an overview of FunConn s two primary toolsets Habitat Modeling and Landscape Networks The Habitat Modeling toolset was designed for those who want to generate a terrestrial habitat quality raster functional patches and a landscape network from the ground up Because the habitat model is based on species vegetation affinities the data requirements are minimal in addition to landcover no existing sampling data are required The Landscape Network toolset is designed for those interested in generating a landscape network based on existing data or analyzing an existing landscape network lt contains three sub toolsets Processing Analysis and Export
43. raph The resource guality threshold is used twice in the Habitat Modeling processes The first use establishes the primary habitat areas from which to base smaller stepping stone habitat areas and ultimately the seeds for defining functional patches While you are setting a threshold for retaining areas of a certain habitat guality areas of lower habitat guality will not be eliminated until their relationship based on distance to the primary patches is evaluated This is done through the patch structure reclass table Minimum Patch Size ha Enter the minimum patch size in hectares to be evaluated This threshold is the smallest biologically relevant patch size for the target organism lt may be based on known home range sizes or by estimating home range size using allometric relationships between body mass and home range size Jetz et al 2004 To ensure that the full range of possible home range sizes is covered we recommend running the model at an order of magnitude less than and greater than the estimated home range size The minimum patch size for the example lynx is 264 ha See Appendix Geoprocessing Steps for details on how this parameter is incorporated into the model Patch Structure Reclass Table This table contains the data necessary to define an organism s response to edge and core habitats For instance if a patch is composed of entirely of high quality land cover type does it decrease in value as the organism approa
44. raster with all cells equal to 1 This will constrain the area of the landscape network Output Landscape Network Enter the name and location of the output landscape network Page 30 7 26 2006 FunConn vl Users Guide Polygons to Landscape Network Tool This tool generates a landscape network based on existing polygon feature class Polygons can be tessellated contiguous or disjoint not share boundaries Examples of possible polygons include the lower 48 states counties or other management units Please note that ArcGIS can place locks on files that are difficult to detect and will prevent proper execution of FunConn If you experience this problem try restarting ArcMap and refreshing your data amp Polygons to Landscape Network DER o Polygon Feature Class Cost Raster optional Resolution Map Units 250 Network Grouping Field Output Landscape Network Cancel Environments Show Help gt gt Parameters Polygon Feature Class Existing polygon feature class Polygons can be tessellated contiguous or disjoint not share boundaries Examples of possible polygons include the lower 48 states counties or other management units Cost Raster Optional Cost raster with continuous values No specific data range is required This is a typical cost surface such that cells that are harder to traverse have smaller values For example if a slope of 65 degrees is nearly impossible
45. s Landscape Network A Landscape Network is a type of graph that recognizes spatial context and relationships with additional geographic information The data structure of a Landscape Network is stored within a ESRI personal geodatabase Landscape Networks have four distinguishing features Theobald 2005 1 The Landscape Network stores both the topology of a graph and the geometry of the nodes and edges possibly multi edges 2 Nodes represent functionally defined patches that represent an organism s behavioral response to landscape structure size shape guality directionality between nodes 3 Effective distance is an attribute of an edge multiple pathways can be unique edges 4 Planar graph algorithms allow for important responses to be modeled such as an organism s use of stepping stones while moving between primary patches Edge While the two definitions for edge are similar the second is specific to the FunConn Tools 1 In graph theory edges connect adjacent nodes See graph definition Page 5 7 26 2006 FunConn vl Users Guide 2 The edges generated by the FunConn Tools are stored within the Landscape Network connect nodes centroid to centroid are not straight line and have a one to many relationship with node pairs multi edge Also each edge is represented twice in the relationship table to account for directionality The Landscape Network edges contain the following attributes edge leng
46. s that fall in the n percentile For example if you specify gro the values falling in the lowest 10 percentile or 10 quantile will be used to assemble the initial linkages between patches This choice will change the number of linkages generated between patches however it can be counter intuitive That is more linkages might result from using a value of gio than a value of q Here s why From the patches allocation zones are grown outward across the cost surface until the meet The area where zones meet is the allocation boundary which is actually 2 cells wide Each boundary cell location has a cost value associated with it Collectively they form a distribution of cost values along the allocation boundary gro is the lowest 10 of the cost values 92 is the lowest 20 of cost values and even though ga includes the go cells fewer groups might be formed after regiongrouping Figure 5 V allocation boundary O I I E Qio cell groups E Q cell groups Q 5 linkages Q 3 linkages Figure 5 Effect of q choice on initial linkage results Output Landscape Network Enter the name and location of the output landscape network The Landscape Network can be analyzed using the Landscape Network Analysis Toolset p 2 Page 27 7 26 2006 Example Lynx Landscape Network Vail Pass Colorado E RE a la za
47. s will be retained Examples of point features include species sampling points plot centroids or management unit centroid Please note that ArcGIS can place locks on files that are difficult to detect and will prevent proper execution of FunConn If you experience this problem try restarting ArcMap and refreshing your data amp Points to Landscape Network o Point Feature Class Resolution Map Units 250 Cost Surface Raster Output Landscape Network Cancel Environments Show Help gt gt Parameters Point Feature Class Existing point feature class When the existing point features are converted to raster format they will default to the cell size of your cost raster Resolution Map Units Enter the correct raster cell size it should not be larger than smallest mapping unit Cost Surface Raster Cost raster with continuous values No specific data range is required This is a typical cost surface such that cells that are harder to traverse have smaller values For example if a slope of 65 degrees is nearly impossible to traverse it might have a cost of 0 01 whereas a slope of 2 degrees would be very easy to move across and have a value of 1 Therefore it would take a much greater number of high cost cells to reach a desired distance than if you were traveling across low cost cells If you do not want to incorporate a cost and generate edges based purely on Euclidean distance still supply a
48. th effective distances centroid to centroid angle and mean angle vector Node In a graph a node is a point functionally connected to other points via edges Nodes are stored in the Landscape Network as a point feature class However patches defined by a polygon feature class can also serve as nodes in the FunConn Landscape Network Analysis tools Linkage Linkages are the least cost pathways between patch edges defined by cost allocation boundaries and a certain threshold that allows for multiple linkages to be defined This threshold is set through the g value which is user defined in the Create Landscape Network tool Also each linkage is represented twice in the relationship table to account for directionality Patch A habitat area functionally defined by habitat guality size and proximity constraints In a traditional graph patch centers serve as the nodes connected by straight line edges In a Landscape Network the patches are stored as a polygon feature class and linkages originate at the patch perimeter Corridor A representation of the optimal movement pathway between adjacent habitat patches Corridors have a one to many relationship between node pairs one corridor can connect several patches The geometry of the corridors reveals potential geographic bottlenecks or other shape characteristics that might enhance or inhibit the traversability of a habitat network Cluster A group of patches that function as a single p
49. to traverse it might have a cost of 0 01 whereas a slope of 2 degrees would be very easy to move across and have a value of 1 Therefore it would take a much greater number of high cost cells to reach a desired distance than if you were traveling across low cost cells Resolution Map Units Enter the correct raster cell size should not be larger than smallest mapping unit Output Landscape Network Enter the name and location of the output landscape network Page 31 7 26 2006 FunConn vl Users Guide Polyline to Landscape Network Tool This tool creates a landscape network from a polyline shapefile or feature class The output landscape network will consist of two feature classes nodes and edges and a relationship table that functionally connects the nodes and edges feature classes The nodes feature class represents the end points of the polylines of the input shapefile with the edges feature class representing the polylines between the end points For this reason it is important to have all line segments in the input polyline shapefile to be snapped to the end points of other line segments since this tool calculates node edge connectivity by geographic coincidence of the end points If the line segments are not snapped to other line segment end points the resulting landscape network will have nodes that are only connected to a single edge instead of nodes that are connected to many edges One way to begin this process
50. us Enter the patch radius in the same units as the map This user defined parameter is the distance that an animal moves on the landscape seeking out forage and is influenced by the organism s perceptual ability Map units are typically in meters Core Habitat Percentage This value is multiplied by the area of the foraging radius n r For lynx it is 0 1 n 9172 Any high quality habitat area that is larger than this area is retained as core seeds trom which to grow functional patches The default value is 0 1 which unless you are performing advanced modeling you don t need to alter Increasing this value will reduce the number of seeds by increasing the minimum size requirement Resource Quality Threshold The resource quality threshold is the minimum habitat quality value acceptable to the target organism to define patches The threshold value will typically fall near 75 80 range 0 100 and is based on the QUALITY values from the Resource Quality Reclass Table The default value is 75 and represents a minimum habitat quality of 75 acceptability to the organism where 100 is the best possible habitat This is the second time the Resource Quality Threshold is used In this tool areas of habitat quality above this threshold are found regiongrouped and then used as the basis for the core seeds from which to grow patches Functional Patches Output Raster Enter the name and location you wish to give this tool s output the Fu
51. y the land cover raster into habitat quality values for each organism Format dbf table Required Fields Case Sensitive 1 GRIDCODE short or long integer type 0 decimal places land cover grid value 2 QUALITY short or long integer type 0 decimal places organism specific value range 0 not habitat 100 optimal habitat quality Example Table c lynx habitat_quality dbf Table 1 Table 1 Example classes from habitat_quality dbf OBID GRIDCODE DESCRIPTION QUALITY 1 2 Rocky Mountain Alpine Bedrock and Scree 0 1 28 Rocky Mountain Subalpine Mesic Spruce Fir Forest and Woodland 100 1 34 Rocky Mountain Ponderosa Pine Woodland 70 1 70 Rocky Mountain Subalpine Mesic Meadow 50 1 96 Chihuahuan Mixed Salt Desert Scrub 0 1 110 Open Water 0 1 112 Developed Medium High Intensity 0 Etc Resource Quality Threshold The resource quality threshold is the minimum habitat quality value acceptable to the target organism to define patches The threshold value will Page 12 7 26 2006 FunConn vl Users Guide typically fall near 75 80 range 0 100 and is based on the OUALITY values from the Resource Ouality Reclass Table The default value is 75 and represents a minimum habitat guality of 75 acceptability to the organism where 100 is the best possible habitat This does not mean that any land cover cell that is below the threshold will be eliminated see next parag
Download Pdf Manuals
Related Search