Visual soil examination and evaluation in the sub-humid and semi-arid regions of Kenya

Soil quality is indicated by the interaction of physical, chemical and biological soil properties. The importance of physical properties, for example soil structure, lies in the fact that they enhance chemical and biological soil functions. Consequently, periodic assessment of structural quality is an important aspect of soil quality management. Quantitative soil properties can be used as indirect indicator parameters for soil structural changes. However, measuring these properties is not applicable, especially for smallholder farmers who cannot afford to pay for laboratory tests. This study contributes to the validation of visual field assessments by comparing the performance of such methods on 'tropical' soils. The study was conducted across two regions with contrasting soil types and land use, in the sub-humid with clay Nitiosl and semi-arid with sandy loam Cambisol locations of Kenya. At both locations, visual methods were tested on soils under cropland and under natural forests (NF). Under the cropland, evaluation sites were selected from researcher and farmer managed sites. Visual scores from visual soil assessment (VSA), visual evaluation of soil structure (VESS) and visual evaluation of soil structure using the core (coreVESS) were correlated with soil physical and chemical properties measured in the laboratory. Under the clay Nitisol, absolute values of Pearson r between VSA scores and laboratory measured soil properties ranged from 0.84 to 0.54, for VESS, they varied between 0.75 and 0.37 while for coreVESS, they ranged from 0.84 to 0.60. For the sandy loam Cambisol, absolute Pearson r values between laboratory measured soil properties and VSA scores ranged from 0.83 to 0.29, the r values were between 0.88 and 0.45 for VESS and between 0.81 and 0.40 for coreVESS. From the obtained correlations, we concluded that the visual methods tested are capable of distinguishing structural quality due to different land use and are therefore suitable for assessing soil structural quality of tropical soils in Africa. Management thresholds were determined using bulk density (BD). The target value for good soil quality (Sq<2) for the Kibugu Nitisol was BD = 0.0012*SOC+0.6476 (r = 0.71; SOC is soil organic carbon), while the trigger and remediation values were 0.93 Mg m- 3 (Sq = 2) and 0.99 Mg m- 3 (Sq = 3), respectively. In the absence of SOC data, the target mean BD for Sq