Groundwater modelling simulates groundwater flow processes underground. Developing useful numerical models requires a range of skills including an understanding of computer code, geology, geographic information systems, databases, statistics and mathematics. We have a team of six hydrogeologists and engineers with these unique skills, who work as a team, constructing numerical models to assist our clients with groundwater management challenges. Our modelling group each brings their own specialities to the team such as uncertainty analysis, calibration techniques, and solute transport modelling.
At AGE we are also proficient in the use of SEEP/W (two-dimensional model) and FEFLOW (finite element three dimensional model) and where necessary can apply this software to meet our clients needs.
How groundwater interacts with surface water is often an important process that needs representing within numerical models. We replicate these interactions using recharge, evapotranspiration and stream flow routing packages using inputs from analytical calculations and other models such as AWBM and SWAT.
Our goal when developing numerical models is to ensure they replicate key underground processes as closely as possible. Numerical models have buyantibioticsonline many parameters that represent the groundwater system, and finding the optimal combination of these parameters can be a challenge.
At AGE we use automated optimisation software such as PEST to improve the calibration of our models. We also have our own high performance computing cluster that speeds up the calibration process allowing many models to run simultaneously. We also employ advanced optimisation techniques including Advanced Spatial Parameterisation and Shuffle Complex Evolution methods when required to ensure each model is well calibrated and suitable for use.
In recent years we have invested heavily in improving our capability to assess the level of uncertainty in the predictions from our numerical models. From this effort we have a growing and advanced capability, supported by a team that includes post-doctoral specialists in numerical uncertainty modelling.
We undertake ‘Monte Carlo’ style analysis that can provide upper and lower bounds on model predictions. This is important for our clients decision making process and can indicate if the level of uncertainty is a risk and whether it requires further work to be reduced. Consequently it can also assess where uncertainty has no material impact on the model predictions.
These types of analyses are made possible through the use of AGEs high performance computing cluster, which is capable of running thousands of model permutations in a matter of hours.