Studying deer habitat on Anticosti Island, Quebec: relating animal occurrences and forest map information

Forest maps are widely available and can provide an inexpensive way to analyse ecological phenomena at a landscape level. This paper presents a case study whose objective was to help manage the deer herd and associated habitat on Anticosti Island. The selection by deer of winter habitat was analysed spatially using aerial survey data and a forest map produced by interpretation of aerial photographs. Initially, it was demonstrated that the spatial distribution of deer across the landscape could not be considered random. Grid cells of two different sizes-500m x 500 m and I km x I km-were extracted from the forest map, and landscape indices thought to be relevant were calculated for each to characterize the key landscape features on which deer select their winter habitat. To do this, the landscape indices were correlated with the number of deer found in grid cells of a given size. It was found that deer preferred areas on which balsam fir was present, and areas in which there was a relatively high concentration of regeneration/dense forest edge. Correlations were better for the larger grid cell size suggesting that deer on Anticosti Island selected habitat based on an area larger than 500 m x 500 m. However, it is noted that this result is probably also due in part to the modifiable areal unit problem (MAUP) whereby larger window sizes tend to provide better correlations between two variables. The location of the grid cells within the study area also affected the results slightly. (C) 2003 Elsevier B.V. All rights reserved.