EROSION by water has been considered as the most serious soil degradation problem in the humid tropical and subtropical India. Dhruvanarayana and Ram Babu analysed the existing soil loss data of the entire country and indicated that soil erosion has been taking place at an average rate of 16.35 t ha–1 yr–1, totalling an annual loss of 5334 million tonnes. Nearly 29% of the total eroded soil is permanently lost to the sea and nearly 10% is deposited in reservoirs, resulting in the reduction of their storage capacity by 1 or 2% annually. The remaining 61% of the eroded soil is transferred from one place to another. Rainfall erosivity, soil erodibility, topography, vegetative cover, management and conservation practices are the major factors affecting soil erosion. Erosion takes place in the form of sheet, rill and gullies. Sheet erosion takes place on slopes due to overland flow. Rills form in the areas where overland flow concentrates. Excess concentration of flow results in gully erosion. On decreasing slopes of overland flow, eroded materials get deposited. Total erosion in the form of sheet, rill and gullies is termed as gross erosion. Sheet and rill erosion from a unit area of a field at a specified slope is defined as soil loss. Singh  et al. reported the annual water erosion rate values from less than 5 t ha–1 yr–1 (for dense forests, snow-clad deserts, and arid regions of western Rajasthan) to more than 80 t ha–1 yr–1in the Shiwalik hills. Ravines along the banks of the rivers Yamuna, Chambal, Mahi, Tapti and Krishna, and the shifting cultivation regions of Orissa and the northeastern states indicated a soil loss exceeding 40 t ha–1 yr–1.

 The annual erosion rate in coastal regions of the Western Ghats representing humid tropical climate varied from 20 to 30 t ha–1 yr–1

 The soils, mainly supporting rainfed agriculture, are subject to severe sheet and rill erosion with an annual soil loss of 20–100 t/ha (ref. 1).  The northeastern states of India represent humid subtropical climate. These states have severe problem of soilwater erosion because of prevalent practices of shifting cultivation (jhumming). In the past, the practice of jhumming worked well because of long fallow cycle (20–30 years). But due to increasing population pressure, the cycle has been narrowed to 3–6 years and thus aggravated the degradation problems due to erosion. Forest tree-cutting, burning, clearing and dibbling of seeds cause nearly 4.1 t ha –1 of soil loss. Soil erosion from hill slopes (up to 60–70%) during the first, second and third years has been reported to be 146.6, 170.2, 30.2 t ha–1 yr –1 respectively .

 The loss of soil due to erosion varies from different land uses in addition to quality of soil and intensity of climatic parameters, especially rainfall and temperature. Since soil erosion is the major reason for soil loss and consequent decline in soil productivity, it becomes imperative for the land-use managers and land-use planners to adopt appropriate soil conservation measures to check this. The type of soil conservation measure again depends on the degree of soil erosion. Thus, it becomes more important to know the amount of soil loss due to erosion. Erosion data have inherent limitations that should be recognized by both resource managers and researchers. Variability, often quite extreme, is characteristic of run-off and soil loss data, and thus sometimes leads to apparently conflicting results. A short-term local plot study does not explicitly evaluate the erosion hazard or treatment effects for an entire landscape. In a one-location study, the effects of rainfall, soil and slope characteristics can rarely be separated because these parameters usually occur at the same levels on all plots in the series or vary in unison. Also, many relevant secondary variables cannot be controlled under natural conditions. Some of these vary randomly over time. Some show seasonal or long-term  trends, but fluctuate unpredictably for short time-periods. The positive and negative effects of random variables and their interrelation thus balance each other over long time-periods. These limitations were alleviated since grid observations (5 km × 5 km) of about 200 numbers in an area of approximately 100,000 ha in Tripura were considered to capture the actual scenario of soil erosion.