Assessing the performance of a plastic optical fibre turbidity sensor for measuring post-fire erosion from plot to catchment scale

Abstract. This study is the first comprehensive testing of a novel plastic optical fibre turbidity sensor with runoff samples collected in the field and, more specifically, with a total of 158 streamflow samples and 925 overland flow samples from a recently burnt forest area in north-central Portugal, collected mainly during the first year after the wildfire, as well as with 56 overland flow samples from a nearby long-unburnt study site. Sediment concentrations differed less between overland flow and streamflow samples than between study sites and, at one study site, between plots with and without effective erosion mitigation treatments. Maximum concentrations ranged from 0.91 to 8.19 g L⁻¹ for the micro-plot overland flow samples from the six burnt sites, from 1.74 to 8.99 g L⁻¹ for the slope-scale overland flow samples from these same sites, and amounted to 4.55 g L⁻¹ for the streamflow samples. Power functions provided (reasonably) good fits to the – expected – relationships of increasing normalized light loss with increasing sediment concentrations for the different sample types from individual study sites. The corresponding adjusted R² values ranged from 0.64 to 0.81 in the case of the micro-plot samples from the six burnt sites, from 0.72 to 0.89 in the case of the slope-scale samples from these same sites, and was 0.85 in the case of the streamflow samples. While the overall performance of the sensor was thus rather satisfactory, the results pointed to the need for scale of site-specific calibrations to maximize the reliability of the predictions of sediment concentration by the POF (plastic optical fibre) sensor. This especially applied to the cases in which sediment concentrations were comparatively low, for example following mulching with forest residues.