Sediment source apportionment using optical property composite signatures in a rural catchment, Brazil

Abstract

Identifying the key sources of fine-grained sediment is essential for protecting and improving soil and water quality. Accordingly, this contribution tested a combination of low-cost analytical procedures for assembling information on key sediment sources in an agricultural catchment in Brazil and in so doing, tested 24 components derived from a conventional printer scanner using various colourimetric models, 18 organic compounds derived from near-infrared (NIR) spectra and 13 soil constituents derived from mid-infrared (MIR) spectra. In combining the application of these low-cost tracers, the study also aimed to investigate potential scale-dependency in sediment sources. Four main sediment sources were sampled: (i) sugarcane; (ii) unpaved roads; (iii) cropland, and; (iv) channel banks and samples were collected from these at two different scales; sub-catchment (~1453 km2) and catchment-wide (~2857 km2). At both scales, channel banks were the most important sediment source followed by sugarcane. At catchment-wide scale, channel banks accounted for 75.9 ± 6.7%, 56.4 ± 16.3%, 39.1 ± 20.7% and 68.3 ± 4.9% of sampled suspended sediment using composite signatures comprising NIR, MIR or colour tracers only, or a combination of all three types of low-cost tracers, respectively. For bed sediment samples, the corresponding respective source contributions were estimated to be 43.4 ± 4.7%, 32.8 ± 7.8%, 49.2 ± 18.6% and 32.0 ± 4.6%. Our results, regardless of optical property, target sediment type, or scale, suggest that channel banks represent the primary source of the sediment problem in the study area. Targeted remedial actions therefore especially need to deliver protection for eroding channel banks.