Two layer packages are included here. One with the final habitat surface and final habitat concentration areas used in the naturalness connectivity models. The other layer package contains an alternate habitat surface and alternate habitat concentration areas that were created in the process of finalizing the model inputs.
Habitat Layer Creation:The habitat layer is needed to create the Habitat Concentration Areas (HCAs) and for use in the linkage mapper model. We wanted the perfect habitat on a landscape to be 1, and the worst to be 0. The habitat layer was created from a “Specialized Resistance Surface” created for this purpose. The “Specialized Resistance Surface” was created with the above methodology, but had a landscape basic resistance of 1 for landscape features such as steep slopes and palustrine forested wetlands. Each of these have energetic costs for movement, but are otherwise wild habitat, and therefore eligible to be HCA for this naturalness based model. (In the standard resistance surface, these landscape characteristics had higher resistance.) (See column F of Resistance_Basemap_v190304_for_HABITAT for the updated values, with yellow highlighting, with notes added in the final column).
From this, we were able to make the specialized Resistance Surface that did not consider protected areas, and the one that did. We then used this equation for each one: (max resistance value on landscape - resistance value of a cell)/max resistance value on the landscape) to yield the two different habitat layers.
Habitat Concentration Areas:Inputs: One of the habitat layers created by one of the resistance layers (see habitat surface metadata). To create the HCAs the Gnarly Tools Core Mapper tool (Shirk and McRae, 2013) was used. The tool uses the Habitat Layer raster, associated Resistance Surface, and user-specified parameters to identify HCAs.We tested different values to become familiar with the parameters and data. We settled on the following values in grey during our first iteration of developing HCAs. Then, we discovered an error in the habitat surface, namely steep slopes were being defined as poor habitat since they are poor for animal movement. We fixed that for the Habitat surface, as per the methods above, and then adjusted the parameters, to the parameter values in black. Parameters:• Moving window radius: (Draft 1,500); final: 1,500• Minimum Average Habitat Value: (Draft 0.92); Final: 0.93 o In the first iteration, we started with a value of 0.9 (a round number and the value of 500-1000 m buffer of abandoned or banked railways), and settled on 0.92. With the new Habitat Layers, we tested this on the Protected Areas habitat layer and found this was yielding a larger proportion of the landscape as core areas than before (41%). We tried 0.93 (37%), 0.94 (33%), and 0.95 (28%). We found this to be a highly sensitive parameter, and found that the runs with 0.94 and higher were missing some areas that we perceived as core areas, both in the standard run and in the protected areas HCAs run. So we went with 0.93..• Minimum Core Area: Draft: 5 square kilometers; Final: 5 square kilometerso size is large enough for small animals to have at least 5, 1 sq km home ranges • Minimum Habitat Value Per Pixel: (Draft: 0.8); Final: 0.8o Thereby allowing small roads to just barely be allowed in HCAs. This avoided giving the HCAs a fragmented look. • Expand Cores by the CWD Value: (Draft: 6,100); Final: 20,000o In the original draft we used 6,100, which was Large enough to link cores separated by small roads, but small enough to minimize the Halo effect created by larger values.o We then used the option in Gnarly Tools to trim back and to exclude non-habitat from core size calculations.o Upon further testing and inspection, we found that increasing this value linked cores separated by a medium road, and found this to be preferable. We tested 10,000 using the protected areas habitat, and this only linked one such core, increasing % protected from 37.49% to 37.54%. Increasing from 6,100 to 20,000, linked 6 such cores, a big improvement, and only increased % protected from 37.49% to 37.77%. Increasing from 6,100 to 30,000, still only linked 6 such cores, and increased % protected from 37.49% to 37.83%. Therefore, we went with 20,000We ran the above parameters for both the Standard Habitat Surface, and also for the Habitat surface with Protected Areas benefits, yielding the two sets of HCAs, WWA_HCAs_Naturalness_Std_3 (N = 73, 32% of the landscape) and WWA_HCAs_Naturalness_PA_6 (N = 73; 38 % of the landscape)We then provided these layers to the experts on the team for review and comment. One suggestion was to start with the PA Layer, then to only use the Gap 1 and 2 areas, where present in really large core areas. This resulted in less area assigned to HCAs. However, they also suggested that the stand alone protected areas that are over 1,000 acres also be added to the HCAs layer. So that was done, and resulted in more area assigned. The result is WWA_HCAs_Naturalness_PA_6_Enhanced (N = 112). This was the layer used for most analyses, aside from some of the sensitivity analyses.
