Mapping continuous forest type variation by means of correlating remotely sensed metrics to canopy N:P ratio in a boreal mixedwood forest

Abstract

AbstractQuestionsCan the ratio of nitrogen to phosphorus (N:P ratio) be predicted at canopy level using imaging spectroscopy (IS) and light detection and ranging (LiDAR) remote sensing data? How do temporal variation and difference in spatial resolution of these data sources affect prediction accuracy of the canopy N:P ratio?LocationBoreal mixedwood forest, northern Ontario, Canada.MethodsCanopy N:P ratio was estimated using spectral indices calculated from IS data at two spatial resolutions, airborne and space‐borne, across two summers. The relationship between the canopy N:P ratio and forest structure was investigated through analysis of LiDAR data. The impact of temporal variation on canopy N:P ratio and the different spatial resolution of IS data on prediction accuracy for canopy N:P was addressed. Maps of canopy N:P ratio generated from airborne and space‐borne IS data were generated.ResultsAirborne and space‐borne IS data explained 70% and 69% of the variation in canopy N:P, ratio, with predictions errors of 5.0% and 7.2%, respectively, in two consecutive years. Predictions differed significantly with changes in spatial resolution. Predictive models obtained from LiDAR data explained 54% and 67% of the variation in canopy N:P ratio, with prediction errors of 6.1% and 7.5%, respectively, for the 2 yrs.ConclusionsThe results show that canopy N:P ratio can be predicted with remote sensing data based on the relationship between canopy N:P ratio and crown closure at this site. The spatial variation due to the mixed deciduous and coniferous forest type is the underlying mechanism that generates the observed spatial pattern in canopy N:P ratio in this ecosystem, and the canopy N:P ratio map displays this variation.