作者: R.A.J. Plant
DOI:
关键词: Evapotranspiration 、 Agricultural land 、 Land use 、 Soil fertility 、 Land use, land-use change and forestry 、 Soil water 、 Inceptisol 、 Soil type 、 Environmental science 、 Hydrology
摘要: Atmospheric concentrations of the greenhouse gas nitrous oxide (N 2 O) have increased significantly since pre-industrial days. Greenhouse gases absorb infrared radiation reflected by earth's surface, thereby causing global warming. The increase in atmospheric N O is attributed to human activities. relative contribution anthropogenic effect about 5%. two major natural sources are soils and oceans, while agricultural comprise main source.Nitrous if formed soil as an intermediate product from nitrification denitrification, processes that operate at microsite scale. Land use changes strongly affect nitrogen (N) cycling; especially conversion forest land generally increases emissions. Therefore, important distal process control on emissions soil. Effects change must be studied scales relevant planning policy making.This calls for methods extrapolate plot-scale measurements highly variable space time. Neglecting spatial heterogeneity fluxes controls can lead serious errors areal regional flux estimates. objective work summarized this thesis was study effects extrapolating Northern Atlantic Zone Costa Rica (2817 ha). A body earlier has been carried out humid tropical region, concurrent availability data soils, use, climate a sizeable area provided unique opportunity in-depth methodological extrapolation. Moreover, history representative regions Latin America.A well-tested process-based ("mechanistic") simulation model driven rainfall events (dndc) used estimate unsampled fields units. units defined such type, management climate. model, originally designed simulate under temperate climatic conditions, adapted justify application pastures banana plantations. First, original formulations evapotranspiration water flow were replaced routines soils. Second, functions simulating i) cattle grazing ii) steady input organic matter through root turnover return excrements pasture added. Third, explicit treatment immobilization added.The tested against field nitric (NO) chronosequence Inceptisols plantation Andisols Inceptisols. For chronosequence, formulation consistent annual dynamics In contrast, simulated daily did not match observations. differences local weather seven sampled sites comprising may caused significant part mismatch. Annual calculated essentially cumulative fluxes, so comparisons provide more conclusive insight model's performance than comparisons.Simulated below compared with monthly frequent sampling. Different parameterizations represent fertilizer inputs plants crop residue additions between plants. both Andisol Inceptisol, below-plant matched frequently measured better fluxes. Simulated between-plant O-N losses Inceptisol 6 15 kg ha -1 , respectively. Field-measured 13 . addition, three fertilization scenarios studied. With fewer equal splits fertilizer-N, loss declined. increasingly depended amount fertilizer-N.An expert system quantifying outputs (pastor) linked produce frequency distributions NO one current ("Natural") alternative systems ("Grass-Legume" "Fertilized Improved"). Current forest-derived deplete stocks therefore unsustainable. Alternative aims utilize soil-N sustainable manner. set up generate parameter sets representing different options systems. rerun each set.Simulated twenty-five years after establishment 3-5 pastures, 12-15 grass-legume mixtures, 7-28 fertilized grasses. NO-N 1-2 7-8 3-16 Regression analysis showed C explained losses, biomass production. Nitrous factor when converted improved pastures. Such sustainability stopping decline N, but necessarily perspective because it emission oxides.The primary secondary forest, plantations explored linking extant Geographic Information System (gis) use. overlaid coverage nearest available meteorological stations. Monte Carlo-based sensitivity identify clay content, initial C, bulk density pH required map attributes key driving variables. 217 units, simulations repeatedly using years.The estimated 1.8-2.1 Gg yr full-fledged performed deterministic stochastic descriptions inputs. accounted across (non-georeferenced) within eleven Using Monte-Carlo integration, obtained per unit class. Regional summing expected values weighted area. Stochastic incorporation variability resulted estimates 14-22% lower those runs, suggesting concavity relationship parameters fluxes.Spatial patterns 1992 evaluated contemporary grasses, (standard deviation parenthesis) 1.0 (0.4) Replacing grasses mixtures groups allowing levels 1.6 (0.5) When all species allowed plantations, 1.9 (1.2) .Land activities terms economic profit fertility unsustainable including extra indicator. Soil variations, dominating patterns, incorporated inventorying Spatial properties regulates finer typically employed surveys. description variables efficient way reduce aggregation estimates.Future challenges include studies conversions, resulting new layouts Also, implicit upscaling potential research