Spatial Ventures expertise offers support in groundwater modelling, steady-state and transient solutions.
Many processes are involved to develop a numerical model based on spatial and temporal finite difference equations requiring data processing, transformation and management.
Model performance criteria depend on a variety of factors, such as discretisation, definition of groundwater related parameters, and the quality, accuracy as well as resolution of real-world representation in the form of spatial data. Solving numerical equations often requires iterative processes, which can be addressed by PEST: A model-independent parameter optimiser used for performing model calibration – the process of estimating model parameters to match known conditions of the aquifer. PEST offers parallel processing, simultaneous model runs can be performed significantly reducing processing/solution time for parameter estimation, especially for transient model calibration. PEST has the ability to consider uncertainty, which is used to describe the potential variability of a model calibration and therefore the uncertainty of the model prediction. It can identify the probabilistic behaviour of the model complexity, which cannot be captured by the calibration process.
We provide groundwater modelling support in formatting and compiling input data (often time-series data, such as rainfall or water-levels), plotting head time-curves (visualising model performance) and statistics describing model prediction performance (e.g. Rmse and Scaled Rmse computed in R), including data analysis (e.g. detecting outliers). We are experienced in working with modelling software packages based on Modflow (2005), such as PMWin and Groundwater Vistas.