Estimating Juniper Cover From National Agriculture Imagery Program (NAIP) Imagery and Evaluating Relationships Between Potential Cover and Environmental Variables

Western juniper (Juniperus occidentalis subsp. occidentalis Hook.) woodlands are expanding from their historic range and causing significant declines of other plant communities. However, landscape-scale restoration projects are hindered by timeconsuming and expensive methods to inventory juniper cover and prioritize landscapes based on developmental phase of juniper encroachment. We investigated the ability of feature-extraction software to estimate western juniper cover from color aerial photographs obtained from the National Agriculture Imagery Program (NAIP) and explored the relationships between juniper cover at stand closure (potential juniper cover) and environmental/site indices (heat load, site exposure, and integrated moisture index) and characteristics measured from commonly available geospatial data layers. Estimates of juniper cover derived from NAIP imagery and ground measurements were similar (R 2 5 0.74; P , 0.01). Neither method consistently estimated juniper cover higher or lower than the other method (P 5 0.79). Environmental indices were either not correlated or weakly correlated with juniper cover at stand closure. However, the environmental/site characteristics (slope, aspect, and elevation) could be used to explain 40% of the variation in juniper cover at stand closure (R 2 5 0.40; P , 0.01). Thus, commonly available geospatial data layers can be used to assist in determining potential juniper cover. This information can then be compared to current juniper cover to determine juniper woodland developmental phase. Knowing the developmental phase is important because management strategies and effectiveness of restoration treatments differ among phases of juniper encroachment. Our results suggest that NAIP imagery can be a valuable tool to estimate juniper cover over large areas effectively to make landscape-scale restoration more feasible. The model of the relationship between environmental/site characteristics measured from commonly available geospatial data layers and potential juniper can be used to assist in restoration planning and prioritization, but could be improved with further refinement.