Nutrient management for rainfed lowland rice in northeast Thailand

Average grain yields of rainfed lowland rice in northeast Thailand are the lowest in the region, and they barely changed in the past decade. Improved fertilizer management is one of the few options to enhance cropping system productivity but related results from previous studies were often disappointing and sometimes contradictory, possibly due to the large variability of soil and water resources. Therefore, the objectives of this study were to develop a site-specific nutrient management approach, and to propose a related decision tool for farmers. For this purpose, we conducted on-farm experiments in Kumpa-Oong Village, Roi Et Province, during the 2003-2005 cropping seasons. Tested fertilizer treatments were designed based on topographic field position. A comparison of soil characteristics between lower and upper fields revealed significantly higher soil fertility for lower fields (higher pH, TOC, TSN, CEC, clay, and silt content; lower sand content). Across seasons and treatments, grain yields were higher in the valley bottom (VB; 2.82 t ha(-1)) than on upper and middle terraces (UMT; 1.68 t ha-1). In all seasons, significant fertilizer treatment effects were detected only in UMT fields. But the comparison of treatment effects in individual fields and in both toposequence positions showed that the limited average fertilizer response was mainly caused by low or even negative responses in fields with a higher control yield, i.e., with higher indigenous nutrient supply. A missing or even negative fertilizer response occurred at lower control yields in UMT fields, most probably because attainable yields in these fields were lower as a result of limited water resources. Thus, site-specific fertilizer recommendations need to take toposequence and the field-specific indigenous nutrient supply into account. Based on these results, we proposed a decision tool that helps farmers to choose the most adequate fertilizer treatment for their fields, based on their knowledge of specific field characteristics. On-farm testing of the proposed decision tool is the next step to show whether this approach is acceptable to farmers and can contribute to higher resource-use efficiency and system productivity. (C) 2009 Elsevier B.V. All rights reserved.