Arizona Missing Linkages
The Arizona Missing Linkages project aims to design wildlife linkages which reconnect important blocks of habitat throughout Arizona. Based on a similar effort in California, our approach creates individual corridor models for 5-15 focal species per study area. These corridors are combined into a multi-species linkage design which - if conserved and integrated with underpasses or overpasses across potential barriers - will best maintain the ability of wildlife to move between protected wildlands even after the remaining land has been converted to uses incompatible with wildlife movement.
In 2005-2006, we created 8 linkage designs throughout Arizona. Because this was the first year of the project, I had many challenges to complete before I could begin analysis, including data acquisition, parameterizing species models, programming analysis tools, and development of a framework to rapidly conduct corridor analyses and create reports. Once a framework was in place, I had to create one 75-150 page report per month on average, so every aspect of the project - modeling, data management, cartography, and writing - had to be optimized for efficiency.
Data management, workflow optimization, and modeling:
Example command line input for corridor analysis. Click for large version.
One of the most challenging aspects of this project was ensuring consistency in modeling procedures across all study areas. To enforce consistency, I wrote a suite of tools in Python to perform all analyses for a species simultaneously (see image). Analyses performed included a series of habitat suitability, patch configuration, and corridor analyses back-to-back for each species, allowing me to batch process all preliminary analyses overnight by supplying just a few key parameters - study area, species name, habitat suitability modeling algorithm, and wildland blocks which we were connecting with corridors. Instead of taking days to construct models by hand, I was able to consistently create them off-hours, allowing me to spend work time concentrating on other tasks.
No modeling study is complete without a couple ulcers, and this project was no exception. While creation of the data management & modeling framework was challenging, it did not make me stress out. What did cause me a little stress was worrying about our how well our modeling procedures represented the habitat preferences of our focal species. While species experts parameterized habitat suitability models for our focal species, some species, such as many herpetofauna, were extremely hard to model using the available GIS data. Given our tight schedule for report production and statewide planning scope, it was impossible to gather new GIS data to create better models. I eventually had to accept that it's not my fault that rattlesnakes like rocky outcrops which aren't available in a GIS layer, and the project was still leaps and bounds above ad-hoc methods for designing corridors.
Cartography:
This project required creation of a large amount of maps (20-50 maps for each report; >200 total maps) in a short amount of time. Land cover maps were the most difficult to create, because they required the display of many classes while maintaining readiblity. ArcGIS's labeling engine and ability to set layer transparencies went a long way towards improving my maps.
Map displaying crossing structures in linkage zone. Click for large version.
Map displaying riparian systems and linkage design. Click for large version.
Map displaying alternative highway realignments. Click for large version.
Map displaying habitat suitability model for mountain lion. Click for large version.
Map displaying important riparian systems. Click for large version.
Map displaying crossing structures and road mortality. Click for large version.
Report writing
Besides programming Python tools, parameterizing species habitat models, conducting analyses, and making maps, I was also the lead writer on every report. We used styles templates in MS Word to manage each 75-150 page report. Although this was successful, Word also successfuly drove me crazy in the process. If I were to write an additional 900 pages of reports in the future, I would strongly consider using LaTeX instead.
Field database storage
We conducted field investigations to document existing crossing structures and recent changes in the landscape. To store our field investigation data for every study area, I created a hybrid geodatabase-MS Access database. Using Python within ArcGIS, the geodatabase imports the coordinates of GPS waypoints, detects all photographs taken at each waypoint, and populates a table linked to each waypoint with metadata such as the date of observation, names of observers, and name of study area. The database is then opened in MS Access, where the user enters information relevant to each waypoint and photo into two different forms. Finally, a report can be generated which shows spatial information, an overview map, and photo notes. As shown in the lowermost photo, each waypoint page in the database report can display up to 4 photos and notes taken from the waypoint. If less than 4 photos were taken for a waypoint, the database uses VBA scripts to adjust the display accordingly.
ArcToolbox database interface. Click for large version.
MS Access forms. Click for large version.
MS Access report created from tool. Click for large version.