Soil Sampling Strategies For The Characterization Of Spatial Variability Under Two Distinct Land Uses

Spatial heterogeneity is a universal phenomenon existing in ecological systems at all scales. In the humid tropics, soil sampling and characterization are complex due to high spatial variability. The inability of researchers to capture this spatial variability inhibits the effective management of resources in terrestrial ecosystems. Hence, simple sampling approaches that are labor and cost-efficient but generate representative samples are required to optimize soil sampling and characterization. In this paper, we used a combination of descriptive statistics and geostatistics to characterize the spatial variability of soil under two land uses. One-way ANOVA and a Student t-test were also used to assess the effectiveness of three distinct sampling designs at capturing the variation in the soil's physicochemical properties. The wetland soils generally exhibited low spatial variability, whereas the variability of the agroforest ranged from low (coefficient of variance (CV) < 25%) to highly variable (CV > 75%). Results show that in the wetland, CVs of samples obtained using the random sampling design (RSD) and response surface directed design (RSDD) were highly correlated (pearson correlation coefficient (r) = 0.98). Similarly, in the agroforest, the CVs of the stratified random sampling design (SRSD) and RSDD gave almost the same accuracy (r = 0.97). Thus, it is recommended that in this and similar humid-tropical wetland and agroforest ecosystems, the RSD and the RSDD, which require the least samples, can be used, respectively. However, more comprehensive testing is required on a broader range of soils before its more widespread application in other climatic conditions.