Resource information
Modern ecological models often account for the influence of the surrounding environment by using landscape pattern indices (LPIs) as measures of landscape structure. Ideally, the landscape samples from which these LPIs are extracted should be centered on the locations where the response variable was measured. However, in situations where this is not possible due to a lack of adequate full-coverage landcover data, the question arises as to what degree this circumstance creates a bias in the value of the LPIs, thereby obscuring their relation with the response variable. To address this question, we extracted four representative LPIs from 30 rectangular (3 × 6 km) landscape samples evenly distributed across a 10 000 km2 boreal forest study area. These rectangles were subjected to systematic displacements across a range of distances (0.5 to 2.5 km) and directions, after which we recomputed the LPIs. We found that a 1 km spatial offset led to an average of 15% deviation of original LPI values. Unfortunately, as the offset increased, the range of resulting deviations also widened, making it difficult to predict this effect. Our findings fill a gap in the literature on landscape pattern analysis and suggest that researchers should avoid LPIs extracted from spatially offset landscape samples.
