Habitat Change Analysis Project
4 likes•1,766 views
The document outlines a GIS-based analysis of wildlife habitat changes in New Jersey from 1986 to 2012, highlighting rapid urbanization's impact, including significant habitat loss and fragmentation. It presents a detailed programmatic approach for assessing habitat metrics and species-specific changes over four time periods, which informs conservation planning and policy-making. The analysis utilizes advanced data management techniques, including PostgreSQL and PostGIS, to track habitat changes and support interactive reporting tools.
Habitat Change Analysis Project
- 1. A programmatic GIS approach to analyzing wildlife habitat change in New Jersey NEARC 2015 Prepared by: Patrick Woerner, GIS Specialist, NJ Division of Fish and Wildlife John Reiser, GISP, Business Intelligence Analyst, Rowan University Sharon Petzinger, Senior Zoologist, NJ Division of Fish and Wildlife
- 2. • Rapid Urbanization/Suburban Sprawl* • Roughly ~15,000 acres per year • Rate of sprawl development gained momentum (~7% increase) over last two decades (while less land available) • Urban surpassed upland forest as dominant land type as of 2007 • Increased impervious surface by nearly nine football fields per day (2002-‐2007) • Myth of Population Growth as Driver • Residential land grew nearly twice as fast as population during 1986-‐2007 period (4x the rate of population growth in the 2002-‐2007 period)* • Habitat Loss • Habitat Destruction • Habitat Fragmentation (loss of habitat functionality) NJ Landscape Context * Hasse and Lathrop (2010) Changing Landscapes in the Garden State: Urban Growth and Open Space Loss in NJ 1986 thru 2007.
- 3. NJDEP Land Use/Land Cover Data (LULC) • Statewide aerial photo interpreted • Modified Anderson (USGS) Classification System • Hierarchical, 86 unique codes • Available for 1986, 1995, 2002, 2007, 2012,… 2015? • Multiple uses, but intended as a resource for change analysis Landscape Project • Habitat mapping for E, T, SC wildlife based on occurrences and LULC-‐derived habitat data • Associates each species with specific set of LULC classes according to habitat requirements • Used for conservation planning, environmental review, habitat management , acquisitions, land use regulation Urban Growth and Open Space Loss in NJ 1986-‐2007 • Ongoing studies based on LULC examining NJ urban growth and land use change • Provides “report card” on urban growth and open space loss looking at time periods 1986-‐95 (t1), 1995-‐02 (t2) 2002-‐07 (t3) and 2007-‐12 (t4) • General reporting on LULC categories used to inform policy Basis of HCAP
- 4. • Tracking of habitat loss and fragmentation, the two greatest threats to wildlife populations • Satisfies State Wildlife Action Plan conservation objectives of evaluating species-‐specific and regional habitat change every five years and assessing trends in loss and conversion • Baseline component for development of species status assessments and recovery plans and use in Delphi Status Review process • Tool to guide and monitor effectiveness of habitat conservation planning, land-‐use regulation and planning, land management, restoration and preservation efforts Overview & Applications
- 5. • Programmatic approach to analysis to obtain multi-‐level estimates of habitat change • Covers four time periods, spanning nearly three decades (T1: 1986 – 1995, T2: 1995 – 2002, T3: 2002 – 2007 and T4: 2007-‐2012) • Incorporates range extents for 60 species, across five taxon (birds, mammals, reptiles, amphibians, and invertebrates) Overview & Applications Common Name Allegheny Woodrat Least Tern American Bi6ern Loggerhead Shrike American Kestrel Long-‐eared Owl Arogos Skipper Longtail Salamander Bald Eagle Mitchell's Satyr Banner Clubtail Northeastern Beach Tiger Beetle Barred Owl Northern Goshawk Black-‐crowned Night-‐heron Northern Harrier Black Rail Northern Pine Snake Black Skimmer Osprey Blue-‐spo6ed Salamander Peregrine Falcon Bobcat Pied-‐billed Grebe Bobolink Pine Barrens Treefrog Bog Turtle Piping Plover Bronze Copper Red-‐headed Woodpecker Brook Snaketail Red-‐shouldered Hawk Ca6le Egret Red Knot Checkered White Robust Baske6ail Cope's Gray Treefrog Roseate Tern Corn Snake Savannah Sparrow Eastern Tiger Salamander Sedge Wren Frosted Elfin Short-‐eared Owl Golden-‐winged Warbler Silver-‐bordered FriNllary Grasshopper Sparrow Superb Jewelwing Gray Petaltail Timber Ra6lesnake Harpoon Clubtail Upland Sandpiper Henslow's Sparrow Vesper Sparrow Horned Lark Wood Turtle Indiana Bat Yellow-‐crowned Night-‐heron Kennedy's Emerald
- 6. Overview & Applications Nuanced, multi-‐dimension species-‐ and habitat-‐ specific change metrics • Species-‐feature label specific (e.g. nesting vs. foraging) • Not only loss/gain/net change, but also transitions between different habitat categories • Fragmentation analysis -‐ number of patches, average, median, minimum, maximum patch size and average, median, minimum, maximum edge-‐to-‐area ratio • % change in habitat category in relation to total area of all habitat (all categories) • % change in habitat category in relation to all change to non-‐habitat • % change of a habitat category in relation to total acreage of that category • Secondary Analysis of WMAs, preservation areas, regulated areas…
- 7. • To form basis of comparative analysis, base layers created following Landscape Project method using LULC from: • 1986 • 1995 • 2002 • 2007 • 2012 Data Development LANDSCAPE BASE LAYER DATA DEVELOPMENT LULC Major Roads Landscape Base Layer Water Buffer (100m) Riparian Clipped by Combined with Erased by || Flood Prone Hydric Soils Wetlands Water Buffer (50m)
- 8. • Species-‐habitat associations derived from the Landscape Project for each unique species-‐feature label (type of occurrence) combination • Habitat selections modified to meet purpose of change analysis • Species-‐habitat associations based on: • peer-‐reviewed scientific literature • occurrence-‐land use analysis to determine preferential selection of certain habitats (i.e., LULC codes used disproportionally to their availability within a species range) • ENSP research and expert opinion Data Development
- 9. Attribute Descriptions Data Development Field Name Description biopid Internal (ENSP) identification code used for individual species spcid Another internal (ENSP) identification code used for individual species spccommonn Common name of species lusort There are 94 lucodes for each species. This number sorts each lucode for each species. lucode NJDEP modified Anderson system land use/land cover (lulc) code. For more information: http://www.state.nj.us/dep/gis/digidownload/metadata/lulc07/ anderson2007.html lupick whether or not the corresponding lucode was considered to be “habitat” for a given species. type broad category of lucode label specific category of lucode rip_only if contains “YES”, this signifies for that specific lucode, only the area that falls within the riparian layer were selected. patchrules Not implemented in this version of the HCAP size_req Patch size thresholds for specified species core_req “Core” requirement for specified species habcat which habitat category the lucode was categorized as for that given species
- 10. • For reporting purposes 96 Anderson codes grouped into 18 habitat categories (habcats) Data Development
- 11. • Species range extents built by generating minimum convex hull on occurrence area data and by incorporating biologists' feedback • Road-‐bound blocks used as consistent units of analysis across all species Data Development Timber Rattlesnake Range Extent and Road Blocks
- 12. Any polygons matching the location criteria and the classification criteria are included and coded for presence or absence in any time period. Data Development
- 13. Top level view of overall habitat impacts – gains and losses Habitat Change Legend Transitional GAIN LOSS Stable Habitat
- 14. Detailed change based on time of habitat change Habitat Change Legend Transitional GAIN T1 GAIN T2 GAIN T3 GAIN T4 LOSS T1 LOSS T2 LOSS T3 LOSS T4 Stable Habitat
- 15. • Species range extents built by generating minimum convex hull on occurrence area data and by incorporating biologists' feedback Golden-‐winged Warbler Golden-‐Winged Warbler Range
- 16. Suitable habitat selected based on range extent and matching land use codes. Golden-‐winged Warbler
- 17. Top level view of overall habitat impacts – gains and losses Golden-‐winged Warbler Legend Transitional GAIN LOSS Stable Habitat
- 18. Detailed change based on time of habitat change Golden-‐winged Warbler Legend Transitional GAIN T1 GAIN T2 GAIN T3 GAIN T4 LOSS T1 LOSS T2 LOSS T3 LOSS T4 Stable Habitat
- 19. -‐12,000 -‐10,000 -‐8,000 -‐6,000 -‐4,000 -‐2,000 0 2,000 4,000 Net 1986-‐1995 Net 1995-‐2002 Net 2002-‐2007 Net 2007-‐2012 Net 1986-‐2012 Timeframe Net Acres Net Changes (acres) in Golden-‐winged Warbler Habitat: 1986-‐2012 Golden-‐winged Warbler
- 20. -‐20,000 -‐15,000 -‐10,000 -‐5,000 0 5,000 10,000 15,000 20,000 25,000 Net 1986-‐1995 Net 1995-‐2002 Net 2002-‐2007 Net 2007-‐2012 Net 1986-‐2012 Net Acres Timeframe Net Changes (acres) in Golden-‐winged Warbler Non-‐habitat Types NON-‐HABITAT AGRICULTURE NON-‐HABITAT NATURAL NON-‐HABITAT URBAN HABITAT Golden-‐winged Warbler
- 21. Non-‐habitat Urban y = -‐2592.9x + 12122 R² = 0.9394 Habitat y = 1977.7x -‐ 7426.2 R² = 0.4341 -‐8,000 -‐6,000 -‐4,000 -‐2,000 0 2,000 4,000 6,000 8,000 10,000 12,000 Net 1986-‐1995 Net 1995-‐2002 Net 2002-‐2007 Net 2007-‐2012 Net Acres Timeframe Net Changes (acres) in Golden-‐winged Warbler Non-‐habitat Types NON-‐HABITAT AGRICULTURE NON-‐HABITAT NATURAL NON-‐HABITAT URBAN HABITAT Golden-‐winged Warbler
- 22. -‐15,000 -‐10,000 -‐5,000 0 5,000 10,000 15,000 20,000 Net 1986-‐1995 Net 1995-‐2002 Net 2002-‐2007 Net 2007-‐2012 Net Acres Timeframe Net Change in Golden-‐winged Warbler Habitat Types SHRUB UPLAND SHRUB WETLAND UPLAND FOREST CON UPLAND FOREST DEC UPLAND FOREST MIX WETEMERG WETLAND FOREST CON WETLAND FOREST DEC WETLAND FOREST MIX Golden-‐winged Warbler
- 23. -‐20,000 -‐15,000 -‐10,000 -‐5,000 0 5,000 10,000 15,000 20,000 25,000 Net 1986-‐1995 Net 1995-‐2002 Net 2002-‐2007 Net 2007-‐2012 Net 1986-‐2012 Net Acres Timeframe Net Change in Golden-‐winged Warbler Habitat Types HABITAT PRIMARY GWWA HABITAT SECONDARY GWWA NON-‐HABITAT AGRICULTURE NON-‐HABITAT NATURAL NON-‐HABITAT URBAN Golden-‐winged Warbler
- 24. Statewide E&T Habitat Change -‐150000 -‐100000 -‐50000 0 50000 100000 150000 200000 T1 T2 T3 T4 Acres Time Period GAIN LOSS Net Change Annualized Change 11,740 13,839 11,919 4,928
- 25. Gloucester County Harrison Township
- 26. Morris County Washington Township
- 27. • Data was prepped using ArcGIS • Union LULC layers to create base data with Anderson Level IV codes for five time periods. • Eliminate sliver polygons • Data was loaded into PostgreSQL • Custom SQL functions perform the selections and filtering necessary • Database views provide for easy reporting and allow for access using ArcGIS software Data-‐Driven Analysis
- 28. • PL/pgsql functions perform selections and spatial functions to produce individual species’ habitat layers. • Polygons are selected, a bitmask is calculated for presence of habitat within a time period, and a view is created to make an ArcGIS layer. Habitat Selection Process
- 29. • A bitmask was employed to accurately and concisely store the habitat status for a given polygon for a given species. • Allows for quick selection of habitat meeting certain time periods. • Allows for easy change to species habitat status • Riparian-‐specific habitat • Core/patch size requirements Using a Bitmask
- 30. • Some species have additional constraints, such as: • Certain land uses must be within a riparian zone to be considered habitat • Land use patches must exceed a certain size • “Core” (inward buffering) of a habitat patch must exceed a size threshold • PL/pgsql functions handle these constraints in additional passes over the data. • Wherever possible, simple value comparisons are performed instead of spatial comparisons, which are expensive. • Riparian is a “precompiled” flag for each base polygon • Core threshold function requires spatial analysis – all performed in SQL • SELECT newregionid, period, ST_Multi(ST_MakeValid( ( ST_Dump(ST_Buffer(shape, -295.276)) ).geom )) as shape FROM biopid45_rd • PostGIS has many spatial functions and operators. Additional Habitat Constraints
- 31. • Once all of the habitat layers have been calculated, we can count the number of species that consider a given base land use polygon as potential habitat. • Counts are performed for each time period and can be used to show change in available habitat due to increased development. Species “Richness”
- 33. • Using Tableau with the spatial database enables interactive dashboards to be created for all of the species. • An interactive website with graphs, maps, and other reports planned for mid-‐2016. • Interactive demo of a habitat change dashboard. Interactive Reporting
- 34. • Having this process in PL/pgsql and PostGIS has considerable benefits: • ArcGIS ModelBuilder could not handle multiple iterations (species, time periods). • ArcGIS Desktop was slow to perform individual steps. • Parameters (such as a species’ land use – habitat preferences and patch size requirements) are in tables – no need to modify the SQL. • Time to produce a single habitat feature class ranges from 11 seconds to 7 minutes. • Entire analysis can be recalculated in a matter of hours. • PostgreSQL and PostGIS are free, well supported software projects. • All of the logic is in version control. • Potential drawbacks: • Spatial SQL may be unfamiliar and is a different approach to the data than Desktop GIS analysis. • Need a DBA (or become familiar with PostgreSQL) Spatial Analysis within a Database
- 35. John Reiser Rowan University Analytics, Systems, and Applications [email protected] @johnjreiser Patrick Woerner NJ DEP Endangered Non-‐Game Species Program [email protected] 609 259-‐6967 Questions, comments? Thank you!