Influence of selective logging and sedimentological process on geochemistry of tropical rainforest streams

Abstract

Stream sediment geochemical surveys are well-suited to mineral exploration in tropical rainforests. However, there is evidence that watershed disturbance can increase sediment loads and thereby dilute or even eliminate geochemical anomalies that, under natural conditions, result from the preferential concentration of heavy minerals by streams. To investigate this and other factors influencing background and threshold values in rainforest streams, sediments have been collected from disturbed and undisturbed watersheds on the Kuamut Formation in the Danum Valley, Sabah, Malaysia. In the laboratory a <2 mm portion was wet-screened to obtain the abundance of the 0.424–2 mm, 0.212–0.424 mm, 0.053–0.212 mm and −0.053 mm size fractions, and the magnetic 0.053–0.212 mm fraction. Another portion was wet-sieved to obtain the −0.212 mm fraction for analysis by ICP–ES after a total decomposition using HF–HCl–HNO 3. Mineralogy of the sediments and heavy mineral concentrates was determined by X-ray diffraction. Results show that although average concentrations of many elements are significantly higher in sediments from unlogged versus logged watersheds, the differences do not simply relate to the effects of logging. Correlation coefficients for the complete data set show two main trends: (1) concentrations of Ca–Mg–Ti–Co–Cr–Sr–V increase with increasing abundance of magnetite; and (2) K–Ba correlate with abundance of fine (<53 μm) sediment. However, a scatterplot of Ti vs K (as geochemical surrogates for magnetite vs fines) shows much more complex relations than indicated by the overall correlations. Most importantly, the Ti–K correlations can be negative or positive depending on which drainage basin the sediments are from. Furthermore, trends in concentration of many trace elements follow the Ti–K trends within individual drainage basins. Insofar as all the streams are on a single geological unit, the presence of distinct linear trends in the scatterplots implies that the composition of the sediments is being systematically modified in a variety of ways. The association between magnetite and Ca–Mg–Ti–Co–Cr–Sr–V (and negative correlation with K–Ba) is best developed in the two largest streams and most likely results from fluvial processes that concentrate heavy minerals and eliminate fines from the stream bed. Conversely, field observations suggest that a trend for Ti–K, together with many other elements, to be positively correlated at low concentrations of Ti depends on whether the sediment source is dominated by bedrock or by regolith supplied from recent mass-wasting events, the frequency of which is increased by logging. It is concluded that selective logging of lowland tropical rainforest does not cause major changes in stream sediment geochemistry except in the vicinity of recent landslides where input of deeply weathered tropical regolith lowers geochemical background. The combined effects of sediment source (landslide regolith versus bedrock) and sedimentological processes on sediment geochemistry can be visualized in X– Y– Z `process' plots, where X– Y are suitable index elements and Z is the element of interest.