Evaluating the effect of data-richness and model complexity in the prediction of coastal sediment loading in Solomon Islands

Abstract

Global biophysical data are increasingly accessible due to improvements in remote sensing and open datasets. These datasets can be of particular value in remote and data-poor environments to enable estimates of water quality impacts from catchment land clearing. Given the resources required to collect field observations and calibrate detailed process-based models, global datasets are often the only sources available to parameterise simple models however the comparative use of these data sources in process-based models is relatively unexplored. This study compares the widely applied models of Integrated Valuation of Ecosystem Services and Trade-offs and Soil and Water Assessment Tool (SWAT) to a tropical catchment in the Solomon Islands using globally available data. These uncalibrated models are contrasted with a SWAT model calibrated with measured streamflow and turbidity in the catchment and meteorologically forced by data from a nearby weather station. These catchment models were coupled with models of sedimentation to examine deposition rates in the coastal lagoon adjacent to the catchment. Model validation using measured coastal sedimentation rates demonstrated that simpler modelling approaches (one-dimensional basin sedimentation and two-dimensional sediment extent modelling) were marginally better than more complex approaches (three-dimensional Delft3D) in data-poor conditions. However, investment in local catchment observations significantly improved the accuracy of simulation outputs. This insight can guide decisions about model complexity, data-richness and investment in local environmental monitoring in these challenging environments.

Keywords

• catchment
• coastal
• data-poor
• remote
• tropical
• sedimentation