Title: Integration of Mid-infrared Spectroscopy and Geostatistics in the Assessment of Soil 3 Spatial Variability at Landscape Level 4

Knowledge of soil spatial variability is important in natural resource management, 22 interpolation and soil sampling design, but requires a considerable amount of geo-referenced 23 data. In this study, mid-infrared spectroscopy in combination with spatial analyses tools is 24 being proposed to facilitate landscape evaluation and monitoring. Mid-infrared spectroscopy 25 2 (MIRS) and geostatistics were integrated for evaluating soil spatial structures of three land 26 settlement schemes in Zimbabwe (i.e. communal area, old resettlement and new resettlement; 27 on loamy-sand, sandy-loam and clay soils, respectively). A nested non-aligned design with 28 hierarchical grids of 750, 150 and 30 m resulted in 432 sampling points across all three 29 villages (730-1360 ha). At each point, a composite topsoil sample was taken and analyzed by 30 MIRS. Conventional laboratory analyses on 25-38% of the samples were used for the 31 prediction of concentration values on the remaining samples through the application of MIRS 32-partial least squares regression models. These models were successful (R2>89) for sand, 33 clay, pH, total C and N, exchangeable Ca, Mg and effective CEC; but not for silt, available P 34 and exchangeable K and Al (R2<82). Minimum sample sizes required to accurately estimate 35 the mean of each soil property in each village were calculated. With regard to locations, 36 fewer samples were needed in the new resettlement area than in the other two areas (e.g. 66 37 versus 133-473 samples for estimating soil C at 10% error, respectively); regarding 38 parameters, less samples were needed for estimating pH and sand (i.e. 3-52 versus 27-504 39 samples for the remaining properties, at same error margin). Spatial analyses of soil 40 properties in each village were assessed by constructing standardized isotropic 41 semivariograms, which were usually well described by spherical models. Spatial 42 autocorrelation of most variables was displayed over ranges of 250-695 m. Nugget-to-sill 43 ratios showed that, in general, spatial dependence of soil properties was: new resettlement > 44 old resettlement > communal area; which was potentially attributed to both intrinsic (e.g. 45 texture) and extrinsic (e.g. management) factors. As a new approach, geostatistical analysis 46 was performed using MIRS data directly, after principal component analyses, where the first 47 three components explained 70% of the overall variability. Semivariograms based on these 48 components showed that spatial dependence per village was similar to overall dependence 49 identified from individual soil properties in each area. In fact, …