Validating the effectiveness and sensitivity of two soil quality indices based on natural forest soils under Mediterranean conditions

Soils from natural ecosystems have specific physical, chemical and biochemical properties determined by the conditions in which these soils have developed. These soils that develop without external disturbance reach a balance amongst their properties. Thus, the creation of a model that represents the established balance of different soil properties from stable ecosystems can be used as a soil quality index, thus any perturbation must lead to modifications in this natural balance. Two regression models with soils from undisturbed forest regions in eastern Spain were previously developed, representing the balance between organic carbon and some physical, chemical and biochemical properties. For undisturbed forest soils, the prediction of soil organic carbon (SOCc) with the calibrated models should be similar to the actual value of this variable (SOCa) (SOCc approximate to SOCa). Consequently, the residuals (SOCc - SOCa) should be around 0. On the contrary, disturbance practices cause a disruption in the balance defined between the different properties and SOC. As a consequence, residuals must be < or >0. Furthermore, the more the degree of degradation increases, the more the values of SOCc must differ from the values of SOCa. According to this, two soil quality indices (SQI) were defined, one for each model, by the calculation of the model residuals (SQI = SOCc - SOCa). The SQIs have been applied to different undisturbed forest soils to evaluate their validity. In addition, they have also been applied to severely altered soils, like agricultural soils, and abandoned agricultural fields, to assess the sensibility of this index to perturbations. After applying the soil quality indices to eleven undisturbed forest soils, it has been verified that a balance exists between organic matter content and different physical, chemical and biochemical properties in forest soils from SE Spain, and the proposed calibrated models are capable of reflecting that balance (SQI = 0). Our results confirm that our models are sensitive to soil perturbation, because agricultural and abandoned agricultural soils have shown an imbalance between organic carbon content and the physical, chemical and biochemical properties (SQI > 0). Moreover, soils from abandoned fields showed lower deviation in the natural equilibrium, indicating a recovery of soil quality. (C) 2008 Elsevier Ltd. All rights reserved.