Development and Field-Testing of the DRAINMOD-FOREST Model for Predicting Water, Soil Carbon and Nitrogen Dynamics, and Plant Growth in Drained Forests.

摘要

A fully integrated DRAINMOD-FOREST model has been developed to simulate the hydrology, soil carbon (C) and nitrogen (N) dynamics, and tree growth for drained forest lands under common silvicultural practices. DRAINMOD-FOREST was developed by linking the hydrological model, DRAINMOD, and the C and N dynamics model, DRAINMOD-N II, to a newly developed forest growth model. The forest growth model estimates net primary production using a method based on radiation use efficiency and allocates fixed C using tree species-dependent allometric relationships. It simulates the effects of environmental factors (temperature and soil water) and N availability on tree growth (both C fixation and allocation). The forest growth module predicts C input to the forest floor due to foliage litterfall and C input to the soil due to root turnover. The model is applicable to mixed or uneven-aged forest stands as it accounts for resource (water, nutrient, and light) competition among different tree species. It simulates silvicultural practices such as thinning, pruning, harvesting, regeneration, and fertilization, and predicts their impacts on water, C and N cycling. The hydrologic model has been modified by adding a revised Gash rainfall interception algorithm and the Penman-Monteith equation to simulate water losses through wet canopy evaporation and dry canopy transpiration, respectively. The hydrologic model, C and N cycling model, and forest growth model are fully integrated, making DRAINMOD-FOREST a comprehensive, quasi-process-based, and stand level model. The functions and features of DRAINMOD-FOREST were demonstrated using a set of long-term simulations covering two typical rotations of a managed loblolly pine (Pinus taeda L.) plantation in eastern North Carolina, United States.;The DRAINMOD-FOREST model was evaluated using a long-term experimental data set from an artificially drained Loblolly pine plantation in eastern North Carolina. The model was calibrated using the data collected during 1988-1997 and validated using the 1998-2008 data. Annual and monthly drainage, as well as daily water table fluctuations were accurately predicted. Annual NPP and daily leaf area index (LAI) dynamics predictions were also comparable to field measurements. Predicted temporal changes in the OC pools on forest floor and in soil profile during the simulation period were reasonable compared to published literatures. Both predicted annual and monthly nitrate export were in good agreement with measured nitrate losses via subsurface drainage. Predicted internal N transformations such as net mineralization, nitrification, and denitrification were also reasonable compared to published literature.;The DRAINMOD-FOREST model was further evaluated using 21-year data collected from two intensively managed coastal loblolly pine plantations located in Carteret county of North Carolina, USA. Simulated management practices included controlled drainage and silvicultural practices consist of N fertilizer application, thinning, forest harvesting, site preparation and regeneration. Predicted annual and monthly drainage as well as daily water table depth were in very good agreements with measured values. Predicted C pool dynamics in forest floor and mineral soil reasonably responded to forest managements and climatic conditions. In addition, the model accurately predicted nitrate losses through subsurface drainage on both annual and monthly bases. We verified the validity of predicted hydrological and biogeochemical responses to controlled drainage and silvicultural practices. Special attentions were given to test the validity of the model in predicting hydrological and biogeochemical processes after forest harvesting. The model also reasonably captured alterations of nitrogen transformations processes caused by forest harvesting, such as increased mineralization, nitrification, denitrification rate, and decreased plant uptake.;A module was added to the DRAINMOD-NII model to describe key mechanisms and processes regulating dissolved organic nitrogen (DON) losses from terrestrial ecosystems. DON production rates were empirically linked with pool size of litter pool on forest floor and in forest soil, as well as soil microbial compartments. The Langmuir isotherm was used to quantify the assumed instantaneous equilibrium between DON in solid and aqueous phases. DON transport with groundwater flow is simulated using numerical solutions to the advection-dispersion-reaction equation. We calibrated and validated the modified model using 20 years of water flow and DON loading data measured at the outlets of three forested (loblolly pine plantations) watersheds located in eastern North Carolina, USA. Field-testing results indicated that the model is capable of reproducing DON export dynamics on both annual and monthly basis. The good model performance is most likely attributed to accurate predictions of drainage rates and reasonable quantification of biotic and abiotic controls on DON dynamics. Although there are some uncertainties of assumptions and methods adapted by the model, the relatively accurate predictions of DON loads indicates a good performance of the model given current limitations of our understanding of inherent factors and mechanisms controlling DON dynamics.;To sum up, this study demonstrated that the DRAINMOD-FOREST model can be utilized to comprehensively predict water, soil C and N dynamics, and plant growth in drained forest ecosystems under intensive management practices.