| Abstract: | Streamwater chemistry was monitored for five years in six streams in a paired catchment experiment in Mendolong, Sabah, Malaysia, comparing the effects of different ways to establish forest plantations with Acacia mangium. At the start of the monitoring in 1985 three catchments were covered with selectively logged rain forest (W4-W6) and three with secondary vegetation after forest fire (W1-W3). The treatments were: (1) clearing of secondary vegetation, burning and planting (W1 and W2); (2) clear-felling, crawler tractor extraction, burning and planting (W5); and (3) clear-felling, manual extraction, no burning and planting (W4). W3 and W6, with no treatment, were monitored as control catchments. Reference monitoring at all streams was for two years and was followed by treatments which lasted for nine months before the full establishment of a new vegetation cover. This paper covers monitoring for a further 2.5 years. The soil types of the catchments were Orthic Acrisol in W3, Gleyic Podsol in W6 and a mix of both soil types in the other catchments. The effect of treatments on streamwater chemistry was clear at both base- and stormflows. Concentrations of major plant nutrients (N, P and K.) became positively correlated to streamflow during treatments. The response of leaching from slash at clear-felling was fast and larger from the clear-felling residues (W4 and W5) than the cleared secondary vegetation (W1 and W2). The intense response to burning was more marked. The stormflow period mean nutrient concentrations were approximately 10-fold for N and K and 10-100 fold for P after burning compared with baseflow mean concentrations over the same period. Significant differences in baseflow concentrations in treated streams generally lasted one year for most elements, but elevated concentrations were still detectable after three years. The first large pulse of leaching was related to mineralization after tree-felling and particularly burning. The longer lasting elevated concentrations in baseflow were associated with the loss of weathering products. The amounts of nutrients lost, calculated by regression analysis as the effect of treatment compared with control, were found to be higher with the degree of vegetation killed and with increased soil disturbance. Consequently, normal forestry practices, with crawler tractor extraction and burning before planting, created the largest leaching losses. The total calculated effect of losses in total N, P and K were (i) W1 + W2 0.5, 1.8, 83.9; (ii) W4 0.8, 0.8, 105.6; and (iii) W5 1.3, 1.3, 189.4 kg ha?1 for the period of 33 months during and after treatment. With normal forestry practice using crawler tractors and with burning before planting (W5), the treatment-induced loss of K was equivalent to 86% of the content of easily decomposed parts of the biomass (leaves, twigs, fine roots and ground vegetation) of the old forest, or larger than K removed by harvest. Exhaustion effects of lowered leaching after repeated burning (forest fire and pre-planting fire) was observed for several elements, indicating possible nutrient deficiencies. |
