Soil Sampling Strategies for Precision Agriculture Research under Sahelian Conditions

al., 1997; Stein et al., 1997). Also, stochastic simulations have been used to reproduce the spatial variability that The cost of soil samples to characterize field variability is a key has been modeled from the observations on simulated problem in precision agriculture. This study was conducted to investi- gate whether yield maps can be used to optimize soil sampling for maps (e.g., Gomez-Hernandez and Srivastava, 1990; characterizing soil variables that determine yield variability. Using an Goovaerts, 1997). inexpensive, low-tech scoring technique, yield maps of pearl millet Apart from these applications and adaptations of ex- (Pennisetum glaucum (L.) Br.) were produced for a zero-input farm isting geostatistical tools, precision agriculture poses in Niger. The soil was classified as a typic Plintustalf. The Spatial some more specific challenges to geostatisticians. One Simulated Annealing (SSA) algorithm was used to optimize three of these is the increasing availability of maps of data sampling schemes. Scheme 1 optimized coverage over the whole area. that can be helpful for purposes such as mapping of soil Scheme 2 covered the whole yield range. Scheme 3 covered the low- properties or yield prediction. Examples of such data producing areas. Yield varied from 0 to 2500 kg ha2 1 , measured per are maps of soil tillage resistance (Van Bergeijk and planting hill. Using correlation coefficients, Scheme 2 found significant Goense, 1997), remote sensed imagery (Booltink and correlations between five soil variables and yield. Scheme 1 found only one significant correlation and explained 37% of the variation Verhagen, 1997), and yield maps collected using low- in yield using multivariate regression of yield on soil variables. Scheme tech (Stoorvogel, 1995) or high-tech (Bouma, 1997) ap- 2 explained 70% of the variation in yield. Differences between Scheme proaches. 3 and Scheme 1 proved to be significant for distance to shrubs, relief, Franzen et al. (1998) found that partitioning the area soil pH, and cation-exchange capacity (CEC). We concluded that according to topography might decrease the number shrubs are the main factor influencing millet yield by means of catching of samples needed to characterize the soil variability eroded materials and improving soil fertility. The possibilities of plant- compared with a regular grid. In this study, we combined ing shrubs to improve soil fertility should be investigated. Variograms such partitioning of the area with geostatistical analysis of relief and yield suggested that spatial correlation is largely confined and interpolation. Furthermore, we present an algo- to distances of 3 to 5 m. Since Scheme 2 was most effective in establish- rithm that optimizes spreading of the observations ing soil-yield relationships, we concluded that yield maps can be used to optimize soil sampling. across the study area. Spatial variability of yield and soils is an important aspect of farming in the zero-input subsistence millet farming systems of the Sahelian Zone. Soils are highly