Linking spatio-temporal variation of crop response with sediment deposition along paddy rice terraces

In tropical mountainous regions of South East Asia, intensive cultivation of annual crops on steep slopes makes the area prone to erosion resulting in decreasing soil fertility. Sediment deposition in the valleys, however, can enhance soil fertility, depending on the quality of the sediments, and influence crop productivity. The aim of the study was to assess (i) the spatio-temporal variation in grain yield along two rice terrace cascades in the uplands of northern Viet Nam, (ii) possible linkage of sediment deposition with the observed variation in grain yield, and (iii) whether spatial variation in soil water or nitrogen availability influenced the obtained yields masking the effect of inherent soil fertility using carbon isotope (13C) discrimination and 15N natural abundance techniques. In order to evaluate the impact of seasonal conditions, fertilizer use and sediment quality on rice performance, 15N and 13C stable isotope compositions of rice leaves and grains taken after harvest were examined and combined with soil fertility information and rice performance using multivariate statistics. The observed grain yields for the non-fertilized fields, averaged over both cascades, accounted for 4.0 ± 1.4 Mg ha−1 and 6.6 ± 2.5 Mg ha−1 in the spring and summer crop, respectively, while for the fertilized fields, grain yields of 6.5 ± 2.1 Mg ha−1 and 6.9 ± 2.1 Mg ha−1 were obtained. In general, the spatial variation of rice grain yield was strongly and significantly linked to sediment induced soil fertility and textural changes, such as soil organic carbon (r 0.34/0.77 for Cascades 1 and 2, respectively) and sand fraction (r −0.88/−0.34). However, the observed seasonal alteration in topsoil quality, due to sediment deposition over two cropping cycles, was not sufficient to fully account for spatial variability in rice productivity. Spatial variability in soil water availability, assessed through 13C discrimination, was mainly present in the spring crop and was linearly related to the distance from the irrigation channel, and overshadowed in Cascade 2 the expected yield trends based on sediment deposition. Although δ15N signatures in plants indicated sufficient N uptake, grain yields were not found to be always significantly influenced by fertilizer application. These results showed the importance of integrating sediment enrichment in paddy fields within soil fertility analysis. Furthermore, where the effect of inherent soil fertility on rice productivity is masked by soil water or nitrogen availability, the use of 13C and 15N stable isotopes and its integration with conventional techniques showed potential to enhance the understanding of the influence of erosion – sedimentation and nutrient fluxes on crop productivity, at toposequence level