Advancing simulations of water fluxes, soil moisture and drought stress by using the LWF-Brook90 hydrological model in R

•Sensitivity analysis of LWF-Brook90 revealed a distinct parameter importance ranking•Bayesian calibration indicated high water use for a poplar short rotation forest•The LWF-Brook90 SVAT model excellently captures temporal variations of soil moisture•The LWFBrook90R package facilitates complex stat...

Full description

Saved in:
Bibliographic Details
Published in:Agricultural and forest meteorology Vol. 291; p. 108023
Main Authors: Schmidt-Walter, Paul, Trotsiuk, Volodymyr, Meusburger, Katrin, Zacios, Martina, Meesenburg, Henning
Format: Journal Article
Language:English
Published: Elsevier B.V 15.09.2020
Subjects:
Bayesian theory
case studies
climate change
Climate impact
data collection
evaporation
evapotranspiration
evergreen forests
grasses
grasslands
hydrologic models
meteorology
model calibration
plant available water
Populus nigra
sensitivity analysis
short rotation
soil depth
soil water storage
water balance
water stress
model calibration
water balance
short rotation
evapotranspiration
Climate impact
sensitivity analysis
ISSN:0168-1923, 1873-2240
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•Sensitivity analysis of LWF-Brook90 revealed a distinct parameter importance ranking•Bayesian calibration indicated high water use for a poplar short rotation forest•The LWF-Brook90 SVAT model excellently captures temporal variations of soil moisture•The LWFBrook90R package facilitates complex statistical analysis and parallelization•The modelling case study with code examples demonstrates the utility of the R package Soil vegetation atmosphere transport (SVAT) models are important for the quantification of water fluxes, soil water availability, drought stress and their uncertainties under climate change. We present LWFBrook90R, an enhanced implementation of the well-established, process-based SVAT model LWF-Brook90 for the R environment for statistical computing. The package provides new functions and sub-models for model parameterization, and facilitates parallel computing, sensitivity analysis and inverse calibration of the model. A case study comprising i) basic forward water balance simulations for temperate grassland vegetation, deciduous and evergreen forest, ii) a parallelized sensitivity analysis, and iii) Bayesian calibrations based on soil water storage observed in a poplar (Populus nigra × P. maximowiczii) Short Rotation Forest (SRF) demonstrates the utility of the R package. The sensitivity analysis revealed parameters affecting plant-available soil water storage capacity and the vegetation's timing and level of water demand to be most important for the annual course of simulated soil water storage, with seasonal and interannual differences in parameter importance rankings. The subsequent calibration yielded a very high agreement between daily simulated and observed soil water storage (0-200 cm soil depth) for the calibration and validation datasets, with Nash-Sutcliffe efficiencies of 0.97 and 0.95, respectively. The final model predicted high though realistic rates of annual evapotranspiration (2011: 844 ± 3.8 mm y-1, 2012: 733 ± 4.5 mm y-1) for the poplar SRF, regularly exceeding grass reference evaporation (ET0) by 20-47% during the months of the growing season. However, basing calibrations solely on observed soil water storage probably resulted in biased partitioning of evapotranspiration towards interception losses. The integration of the LWF-Brook90 hydrological model into R with its wide variety of extensions was successfully tested and may provide efficient, reliable and reproducible water balance predictions by facilitating complex statistical analyses and large-scale applications of the model.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0168-1923
1873-2240
DOI:10.1016/j.agrformet.2020.108023