Evaluation of alternative interpolation techniques for the mapping of remotely-sensed submersed vegetation abundance

New remote sensing technologies have emerged to quantitatively assess submersed aquatic vegetation abundance and distribution. We evaluated a hydroacoustics global positioning system to map the percent of the water column occupied by submersed vegetation (referred to here as biovolume) in three Minnesota (USA) lakes. We evaluated the relative accuracy and precision of digital biovolume maps produced by three interpolation methods (inverse distance weighted (IDW), kriging and spline) after using a non-parametric regression smoother to remove a non-linear depth trend. Interpolated predictions with all methods were relatively accurate in all lakes; however, precision varied among lakes. In all cases, kriging interpolation produced the best predictions when compared with observations in independent verification data sets. However, IDW predictions were only slightly less precise. Map detail was lost when sampling effort was reduced from 10 m transect spacing to 20 or 40 m, although estimates of littoral-wide means did not change appreciably. We concluded that hydroacoustics combined with geostatistics and interpolation in GIS can accurately and precisely display multi-scale patterns in biovolume.