Organic matter restoration by conversion of cultivated land to perennial pasture on three agro-ecosystems in the Free State

English: Understanding the process of organic matter degradation and restoration is important with regard to sustainable agricultural production on any agro-ecosystem, and of particular importance where degradation is relatively rapid, such as in the coarse textured savannah soils of the South African highveld. Organic matter degradation studies on such soils in three agro-ecosystems, Harrismith, Tweespruit and Kroonstad, have been undertaken by Du Toit et al. (1994), and Lobe et al. (2001). This study is concerned with organic matter restoration on the same agro-ecosystems, and is therefore complementary to the two earlier studies. The objective was to investigate organic matter restoration at three depths, 0-50, 50-100 and 100- 200 mm, on perennial pastures of different ages that had been established on lands which had been cultivated continuously for more than 20 years. Representative C and N values for degraded lands and virgin grasslands for the three agro-ecosystems were obtained from the studies of Du Toit et al. (1994) and Lobe et al. (2001), and used as reference values. To reduce within-site error samples were collected at six places, separated from each other by a few meters, at each site. At each of these places six subsamples of each layer were taken to make up the final sample. There were therefore 18 soil samples per site. A total of 28 sites, ranging in ages from 4 to 25 years, were identified and sampled on the three agro-ecosystems, All the samples were analyzed for C and N, and selected samples were analyzed to characterize the soil fertility levels and particle size distribution at each site. Results showed a wide variation in the rate of organic matter restoration between sites in each of the agro-eco systems , due mainly to differences in natural resource factors and management techniques. Most important of the latter was the application of N fertilizer. Where this was inadequate or absent, very low organic matter restoration rates were generally measured. An approximate threshold value of available N below which organic matter restoration is severely impaired appears to be about 15 mg kg". On pastures up to the age of25 years most of the C and N storage has been in the 0-50 mm layer, a little in the 50-100 mm layer, and very little in the 100-200 mm layer. This observation accentuates the importance of the sampling depth in such studies. These results are in accordance with those of Potter ef al. (1999). The mean C gains over all the sites in the three agroecosystems, excluding those with a Nfertility level considered too low to initiate efficient C sequestration, is 0.56 Mg ha-I yr' as compared to 0.8 Mg ha" yr" suggested by Bruce el al. (1999) for the United States of America and Canada. The relatively coarse texture of the Free State soils, and the lower aridity indices, may account for the difference. An attempt was made by pooling the data for the three agro-ecosystems, and adopting a normalization procedure, to identify common C and Nrestoration curves with time. Although a definite upward trend is visible, large inter-site variation and the shortage of data points above 20 years results in relatively low correlation coefficients and the curves being unreliable at their top end. Further research to obtain data from very old pastures is recommended, as well as ecotope specific research on benchmark ecotopes to define in a reliable way the shape of the organic matter restoration curve.