Dataset: GLO MFdmax v01

Description

Abstract

The dataset was derived by the Bioregional Assessment Programme from multiple source datasets. The source datasets are identified in the Lineage field in this metadata statement. The processes undertaken to produce this derived dataset are described in the History field in this metadata statement.

This dataset uses the results of the design of experiment runs of the alluvial MODFLOW groundwater models for the Avon and Karuah alluvial systems of the Gloucester to train emulators to (a) constrain the prior parameter ensembles into the posterior parameter ensembles and to (b) generate the predictive posterior ensembles of maximum drawdown and time to maximum drawdown. This is described in product 2.6.2 Groundwater numerical modelling (Peeters et al. 2016).

Peeters L, Dawes W, Rachakonda P, Pagendam D, Singh R, Pickett T, Frery E, Marvanek S, and McVicar T (2016) Groundwater numerical modelling for the Gloucester subregion. Product 2.6.2 for the Gloucester subregion from the Northern Sydney Basin Bioregional Assessment. Department of the Environment, Bureau of Meteorology, CSIRO and Geoscience Australia, Australia. http://data.bioregionalassessments.gov.au/product/NSB/GLO/2.6.2.

Dataset History

This dataset contains all the scripts to carry out the uncertainty analysis for the maximum drawdown and time to maximum drawdown at the groundwater receptors associated with the Avon and Karuah alluvia and all the resulting posterior predictions. This is described in product 2.6.2 Groundwater numerical modelling (Peeters et al. 2016).

The file 'GLO_MF_dmax_UA_workflow.pptx' shows a diagram of the workflow of the uncertainty analysis for the Avon model. The workflow for the Karuah uncertainty analysis is identical. Avon specific files have 'g' in the filename. Karuah specific files have 'k' in the filename.

The post-processed results from the numerical modelling (dataset GLO MF Model v02) are summarised in files 'GLO_MF_karuah_budget_m3d.csv', 'GLO_MF_gloc_budget_m3d.csv', 'GLO_MF_dmax_g_DoE_Predictions.csv' and ''GLO_MF_dmax_g_DoE_Predictions.csv' .

Python script 'GLO_MF_dmax_Predictions.py' creates input files for the uncertainty analysis:

File 'GLO_DoE_Parameters.csv' contains the 10.000 parameter combinations used in the design of experiment, generated with the maximin latin hypercube sampling.

File 'GLO_MF_dmax_Parameters.csv' contains information on the parameters of the Avon and Karuah MODFLOW model that are varied in the uncertainty analysis. This is the name, transform, minimum, maximum, description, the type of prior distribution, the first and second moments of the prior distribution and the covariance. These values are specified by the modelling team and described in section 2.6.2.6 in Peeters et al. (2016).

The script 'GLO_MF_SensitivityAnalysis.py' calculates the Plischke et al (2013) sensitivity index for all HRV-receptor-parameter combination, which are stored in files 'GLO_MF_dmax_SI.csv', 'GLO_MF_tmax_SI.csv' and 'GLO_MF_sumstat_SI.csv'.

For each prediction of tmax and dmax an emulator is created with R-scripts 'GLO_MF_dmax_g_Emulate_pred_i.R' and 'GLO_MF_dmax_g_Emulate_pred_i.R'. The script uses the R-scripts in dataset 'R-scripts for uncertainty analysis v01' and stores an R data object with extension .Rdata in subdirectory 'emulators/predictionname'.

The R script 'GLO_MF_dmax_CreatePosteriorParameters.R' performs the Approximate Bayesian Computation Markov Chain Monte Carlo to constrain the prior parameter distributions with an acceptance criterion based on the historical water balance of the alluvial system. This model result is summarised in file 'GLO_sumstat.csv' and all results of the design of experiment are stored in 'GLO_sumstat_DoE.csv'. The script results in an emulator object (file '\emulator\glo_sumstat\glo_sumstat.RData') and a spreadsheet with 10.000 posterior parameter combinations (file 'GLO_MF_dmax_g_Posterior_Parameters.csv'). The first 200 of these parameter combinations are used to generate the output for interpolation of exceedance probabilities in dataset GLO MF Model v02.

The R scripts 'GLO_MF_tmax_g_MCsampler_pred_i.R' and 'GLO_MF_dmax_g_MCsampler_pred_i.R' use the emulators for each prediction to sample the posterior parameters to create the posterior predictive ensembles, which are stored for each prediction separately in '\emulators\predname_prediction.csv'

Post-processing scripts 'GLO_MF_dmax_Postprocess_Predictions.py' and 'GLO_MF_tmax_Postprocess_Predictions.py' combine these individual files into two posterior prediction files: 'GLO_MF_dmax_Posterior_Predictions.csv' and 'GLO_MF_tmax_Posterior_Predictions.csv'. These have the combined predictions at the receptor locations of both the Karuah and Avon alluvial MODFLOW models.

These files are further summarised in exceedance probabilities with script 'GLO_MF_probabilities.py', which outputs files 'GLO_MF_dmax_ExceedanceProbabilities.csv' and 'GLO_MF_tmax_ExceedanceProbabilities.csv'

References

Peeters L, Dawes W, Rachakonda P, Pagendam D, Singh R, Pickett T, Frery E, Marvanek S, and McVicar T (2016) Groundwater numerical modelling for the Gloucester subregion. Product 2.6.2 for the Gloucester subregion from the Northern Sydney Basin Bioregional Assessment. Department of the Environment, Bureau of Meteorology, CSIRO and Geoscience Australia, Australia. http://data.bioregionalassessments.gov.au/product/NSB/GLO/2.6.2.

Plischke E, Borgonovo E, and Smith CL (2013) Global sensitivity measures from given data European Journal of Operational Research 226, 536-550

Dataset Citation

Bioregional Assessment Programme (XXXX) GLO MFdmax v01. Bioregional Assessment Derived Dataset. Viewed 18 July 2018, http://data.bioregionalassessments.gov.au/dataset/df0ff3b6-b131-4cf7-9d86-33d1de01634d.

Dataset Ancestors

General Information

Distributions