Point-to-Grid Conversion in Flux Footprints: Implications of Method Choice and Spatial Resolution for Regional-Scale Studies

Abstract

Flux-footprint modelling is commonly used for determining source areas of land–atmosphere gas-exchange observations from a single meteorological tower to determine the relative flux contribution of different areas within the station fetch. In doing so, a conversion from a continuous function to a two-dimensional spatial representation takes place prior to the interpretation of results. This interpreted result is commonly used to validate satellite regional-scale, remote-sensing-based, gas-exchange models. However, little attention has been paid to exactly how the conversion is performed. The conversion to a raster data model can result in the loss of information. This loss can be cell size dependent. Therefore, it is important that the spatial-conversion method used is appropriate for the specific application. We present here an assessment of the influence of spatial resolution (i.e., cell size) and conversion method on the accuracy of modelled flux footprints using tower-based flux instrumentation. Five methods for aggregating flux footprints to a two-dimensional grid are presented and a sensitivity analysis is performed. Two well-validated flux-footprint models are used at spatial resolutions ranging from 10 m to 1000 m. Results indicate significant data loss, change of apparent maximum location, and change in the representative fetch of the point measurement. Furthermore, these results are measurement site dependent. This often-overlooked step in the conversion of a flux-footprint model to a grid is important. These results provide guidelines for those utilizing a flux-footprint model to validate spatial models for assessing gas exchange.