Long term fertility experiments on intensive rice systems in the Philippines: leading indicators of soil fertility and sustainability.

Long-term fertility experiments (LTFEs) were initiated in the mid-to late 1960s at three locations in the Philippines: the International Rice Research Institute (IRRI) [Los Banos, Laguna]; the Philippine Rice Research Institute (PhilRice) [Munoz, Nueva Ecija]; and the Bicol Integrated Agricultural Research Center (BIARC). Treatments consisted of an unfertilized control, +N, + NP, + NK, and + NPK and three best available varieties in a completely randomized block design. From the 1960s to 1991, the N rates at IRRI were 140 kg/ha in the dry season (DS) and 60 kg/ha in the wet season (WS) and the P and K rates were 13 and 25 kg/ha, respectively, for both seasons. At PhilRice and BIARC, N rates were 140 in DS and 70 kg/ha in WS, while P and K rates were 26 and 50 kg/ha, respectively for both seasons. Grain yields were measured in each crop cycle; soil properties and total nutrient uptake and removal were monitored periodically. The effective soil N supply and grain yield decreased in both + NPK and unfertilized treatments, despite conservation or increase in soil organic carbon (SOC) and total N (TN) at IRRI and PhilRice. Mean yield increased (1.1-2.1 t/ha) when N rates were increased from 140 kg/ha (1989-91) to 200 kg/ha in + NPK treatments (1992-94 DS) at all sites. Soil P also increased which was consistent with the partial input/output P balance. However, K removal by harvested portions exceeded fertilizer K inputs by 25-83 kg/ha. The net K deficit was greater in the + NPK plots than in the unfertilized plots at IRRI and PhilRice. Despite the negative partial K balance at IRRI and BIARC, K from irrigation water maintained soil K levels. At PhilRice, soil K levels decreased from 1968-91 in the + NPK treatments when 50 kg K/ha was applied to each crop, hence, in 1992, the K rate was increased to 100 kg/ha to maintain adequate K supply. It was concluded that fertilizer inputs at 20 kg P and 72 kg K/ha are required for each rice crop to sustain soil P and K fertility assuming that yield of 7-8 (DS) and 5-6 t/ha (WS) are achieved, rice stubbles are returned, and P or K inputs from irrigation water are negligible. Although increased N fertilizer rates were required to maintain yields at IRRI and PhilRice, this increase was not associated with a decrease in SOC or TN nor was related to the use of fertilizers. LTFEs are good indicators of system performance and changes in soil fertility in intensive rice cropping.