STATEWIDE ASSESSMENT OF WILD TURKEY HABITAT USING SATELLITE IMAGERY IN ARKANSAS
Abstract: We explored the potential provided by classified digital land cover maps derived from remotely sensed satellite imagery for assessing statewide habitat suitability for eastern wild turkeys (Meleagris gallopavo silvestris) in Arkansas. We adapted habitat variables used for ground‐based habi...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2005-01-01
|
| Series: | Wildlife Society Bulletin |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/j.2328-5540.2005.tb00309.x |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract: We explored the potential provided by classified digital land cover maps derived from remotely sensed satellite imagery for assessing statewide habitat suitability for eastern wild turkeys (Meleagris gallopavo silvestris) in Arkansas. We adapted habitat variables used for ground‐based habitat evaluation to classified land cover and assessed quality of the landscape for turkeys by using 2 approaches: a habitat suitability index (HSI) model and logistic regression model. We acquired digital land cover data derived from satellite imagery from the Multi‐Resolution Land Characteristics (MRLC) consortium, and computed composition and configuration variables with FRAGSTATS (ArcView Patch Analyst). The HSI model incorporated food and cover variables into a geographic information system (GIS; ArcView Spatial Analyst) and evaluated habitat at a pixel resolution of 30 m. We summarized HSI scores at the county level and regressed against harvest records for wild turkeys. HSI values for the statewide model ranged from 0.52 to 0.79 and explained 32% of the variation in harvest (r2 = 0.32, n = 68, P < 0.05). Models tailored to each of 4 regions showed habitat suitability ranged from 0.07 to 0.92 and accounted for nearly 70% of the variation in harvest (Ouachita region; adjusted r2 = 0.68, n = 13, P < 0.05). We used logistic regression to derive a habitat model by comparing land cover characteristics and harvest. We identified 2 variables as most often associated with low harvest of wild turkeys statewide: percentage of land in Row Crops (χ2 = 10.08, df = 1, P < 0.002) and percentage of land in Commercial‐Industrial‐Transportation (χ2 = 8.96, df = 1, P = 0.028). Our findings suggest that NLCD satellite imagery and GIS tools can be used to identify habitat characteristics that allow assessment of the potential of landscapes to support wild turkey harvest. If harvest statistics provide a reasonable surrogate for relative population abundance for wild turkeys, then these models are good indicators of habitat suitability. |
|---|---|
| ISSN: | 2328-5540 |