I'm really curious with how Maxent setting work. I'm using it for elephant habitat mapping. I'm using various variable but I got a problem there. So the first picture is my output, and the second picture the output overlayed by elephants presence. The north (up side of the map) is actually match the data, but the other is actually false since it's city and of course not suitable for elephant. I realized that the problem is because the north side have many elephants presence and the habitat have same characteristic with the middle side so the algorithm count it as a suitable habitat for elephant. Unfortunately the west side that have some elephant points and the south side is actually the one that suitable for habitat, but the algorithm can't read it since the north sample is too dominating. I've tried to delete some of the north sample but it don't make big difference. Please enlighten me about what should I do maybe the Maxent setting or other thing. I'll provide more information about my Maxent setting if it's needed. Thanks!
Tim Lucas unfortunately I don't add any variable that measures urban, I only use environment variable (e.g. slope, NDVI, surface temperature, distance from river, distance from forest). I've tried once using Euclidean distance from settlement but it generate a weird statistic since there are some little settlement like cottage or hut that are located inside the forest so the algorithm make my habitat is close to settlement. Any recommendation for what variable I should add or change?
You might try population density as a variable and consider thinning your dataset so you don't overrepresented the northern areas by selecting only one occurance from a grid of certain size, e. g. 1 or 100 km² depending on the density...
this case solved! I've used two recommendation above to thinning my dataset and use population density. Thank you!
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Invertebrate Zoology
- Arthropods
- Insects