An evaluation of remote sensing derived soil pH and average spring groundwater table for ecological assessments

Abstract

tEcological assessments such as species distribution modelling and benchmarking site quality towardsregulations often rely on full spatial coverage information of site factors such as soil acidity, moistureregime or nutrient availability. To determine if remote sensing (RS) is a viable alternative to traditionaldata sources of site factor estimates, we analysed the accuracy (using ground truth validation measure-ments) of traditional and RS sources of pH and mean spring groundwater level (MSL, in m) estimates.Traditional sources were a soil map and hydrological model. RS estimates were obtained using vegetationindicator values (IVs) from a Dutch national system as an intermediate between site factors and spectralresponse. IVs relate to those site factors that dictate vegetation occurrence, whilst also providing a robustlink to canopy spectra. For pH, the soil map and the RS estimate were nearly as accurate. For MSL, theRS estimates were much closer to the observed groundwater levels than the hydrological model, but theerror margin of the estimates still exceeded the tolerance range of moisture sensitive vegetation. Therelatively high accuracy of the RS estimates was made possible by the availability of local calibrationpoints and large environmental gradients in the study site. In addition, the error composition of the RSestimates could be analysed step-by-step, whereas the traditional sources had to be accepted ‘as-is’. Alsoconsidering that RS offers high spatial and temporal resolution at low costs, RS offered advantages overtraditional sources. This will likely hold true for any other situation where prerequisites of accurate RSestimates have been met.