Evaluation of biogeochemical models at local and regional scale

Additional index words: nutrient cycling, soil modelling, uncertainty analysis, calibration, scenario analysis, model errorIn this thesis different nutrient cycling and soil acidification models, developed for use at different scales, are presented and evaluated. The models considered are Nucsam (NUtrient Cycling and Soil Acidification Model), Resam (REgional Soil Acidification Model) and Smart2 (an extended version of Simulation Model for Acidification's Regional Trends). These are mechanistic dynamic models, which simulate biogeochemical processes in semi-natural terrestrial ecosystems at a variety of scales. The research tool Nucsam, which is specifically developed for application on a local scale, includes simulation of the biogeochemical processes in various soil layers and on a daily time-scale. Resam and Smart2, tools to support policy makers, were specifically developed to evaluate long-term soil responses to deposition scenarios on a regional scale (national to continental, respectively). For that reason, the models Resam and Smart2 are relative simple models and operate on a yearly time-scale. These models were developed in view of the following research hypotheses:Adequate simulation of temporal responses in soil solution chemistry on a daily basis at various depth requires a detailed multi-layer biogeochemical model (Nucsam);Annual average responses in soil solution chemistry at the bottom of the root zone can be adequately simulated with a simple, one-layer biogeochemical model (Smart2);Simulation of soil solution chemistry on a regional scale requires a simplified model;Adequate simulation of soil solution chemistry on a regional scale requires parameterisation, calibration, validation and uncertainty analysis on that scale.Therefore, this thesis primarily aims at testing these hypotheses by (i) validation and calibration, (ii) uncertainty analysis, and (iii) model comparison. More specifically, the models Nucsam (site scale), Resam (site scale/regional scale) and Smart2 (regional scale) will be evaluated with respect to the optimal balance between model complexity, data availability and model aim.The detailed model Nucsam reproduced the magnitude and trends of measured quantities, such as soil water contents and soil solution chemistry, fairly well. However, the application on a site scale hampers from the lack of sufficiently good quality data. A model, such as Nucsam, can not be applied at a large spatial scale because of the lack of data availability. The simplified model Smart2 is capable to simulate the observed flux-weighted annual averaged concentrations. Ignoring seasonal variations of weather conditions, ignoring of different soil layers and simplifying process description simplification does not need to greatly affect the modelled long-term annual average responses to acid deposition. A simplified model, such as Smart2, is an acceptable tool for making long-term evaluation of environmental abatement strategies. Model performance is seriously improved and the prediction uncertainties strongly decreased by model calibration at the scale required for the ultimate output. Further improvement through calibration is hampered from the lack of good quality data on a national scale.

Bibliographic Details

Main Author:	Kros, J.
Other Authors:	van Breemen, N.
Format:	Doctoral thesis biblioteca
Language:	English
Subjects:	acidification, biogeochemistry, calibration, computer simulation, cycling, models, nutrients, soil, biogeochemie, bodem, computersimulatie, kalibratie, kringlopen, modellen, verzuring, voedingsstoffen,
Online Access:	https://research.wur.nl/en/publications/evaluation-of-biogeochemical-models-at-local-and-regional-scale-2
Tags:	Add Tag No Tags, Be the first to tag this record!