RESUMO
Soil organic matter (SOM) is an indicator of sustainable land management as stated in the global indicator framework of the United Nations Sustainable Development Goals (SDG Indicator 15.3.1). Improved forecasting of future changes in SOM is needed to support the development of more sustainable land management under a changing climate. Current models fail to reproduce historical trends in SOM both within and during transition between ecosystems. More realistic spatio-temporal SOM dynamics require inclusion of the recent paradigm shift from SOM recalcitrance as an 'intrinsic property' to SOM persistence as an 'ecosystem interaction'. We present a soil profile, or pedon-explicit, ecosystem-scale framework for data and models of SOM distribution and dynamics which can better represent land use transitions. Ecosystem-scale drivers are integrated with pedon-scale processes in two zones of influence. In the upper vegetation zone, SOM is affected primarily by plant inputs (above- and belowground), climate, microbial activity and physical aggregation and is prone to destabilization. In the lower mineral matrix zone, SOM inputs from the vegetation zone are controlled primarily by mineral phase and chemical interactions, resulting in more favourable conditions for SOM persistence. Vegetation zone boundary conditions vary spatially at landscape scales (vegetation cover) and temporally at decadal scales (climate). Mineral matrix zone boundary conditions vary spatially at landscape scales (geology, topography) but change only slowly. The thicknesses of the two zones and their transport connectivity are dynamic and affected by plant cover, land use practices, climate and feedbacks from current SOM stock in each layer. Using this framework, we identify several areas where greater knowledge is needed to advance the emerging paradigm of SOM dynamics-improved representation of plant-derived carbon inputs, contributions of soil biota to SOM storage and effect of dynamic soil structure on SOM storage-and how this can be combined with robust and efficient soil monitoring.
Assuntos
Ecossistema , Solo , Carbono , Clima , PlantasRESUMO
As agricultural land area increases to feed an expanding global population, soil erosion will likely accelerate, generating unsustainable losses of soil and nutrients. This is critical for Kenya where cropland expansion and nutrient loading from runoff and erosion is contributing to eutrophication of freshwater ecosystems and desertification. We used the Revised Universal Soil Loss Equation (RUSLE) to predict soil erosion rates under present land cover and potential natural vegetation nationally across Kenya. Simulating natural vegetation conditions allows the degree to which erosion rates are elevated under current land use practices to be determined. This methodology exploits new digital soil maps and two vegetation cover maps to model topsoil (top 20 cm) erosion rates, lifespans (the mass of topsoil divided by erosion rate), and lateral nutrient fluxes (nutrient concentration times erosion rate) under both scenarios. We estimated the mean soil erosion rate under current land cover at ~5.5 t ha-1 yr-1, ~3 times the rate estimated for natural vegetation cover (~1.8 t ha-1 yr-1), and equivalent to ~320 Mt yr-1 of topsoil lost nationwide. Under present erosion rates, ~8.8 Mt, ~315 Kt, and ~ 110 Kt of soil organic carbon, nitrogen and phosphorous are lost from soil every year, respectively. Further, 5.3 % of topsoils (~3.1 Mha), including at >25 % of croplands, have short lifespans (<100 years). Additional scenarios were tested that assume combinations of terracing and reduced tillage practices were adopted on croplands to mitigate erosion. Establishing bench terraces with zoned tillage could reduce soil losses by ≥75 %; up to 87.1 t ha-1 yr-1. These reductions are comparable to converting croplands to natural vegetation, demonstrating most agricultural soils can be conserved successfully. Extensive long-term monitoring of croplands with terraces and reduced tillage established is required to verify the efficacy of these agricultural support practices as indicated by our modelling.