Technical Learning Exchange on groundwater data in decision support tools

DHI conducted a technical review and enhancement of the City of Cape Town's Bulk Water Decision Support System (BWDSS). Key deliverables included a custom PostgreSQL timeseries provider to reduce database redundancy, a guide for operationalizing transient FEFLOW groundwater models for the Atlantis, Cape Flats, and Table Mountain aquifers, and recommendations for database maintenance and script optimization to improve system resilience and performance.

• The Bulk Water Decision Support System (BWDSS) used by the City of Cape Town consists of three primary applications: MIKE Operations for backend management with a PostgreSQL database, WaterNet Advisor for near real-time water allocation scenarios based on an EPANET model, and Bulk Water Web Dashboards and Data Capture (Web DSS) for remote data collection and visualization.
• Analysis of the BWDSS database revealed performance bottlenecks caused by large tables, specifically the 'blob' table (3,214 MB), 'change_log' (1,003 MB), and 'time_series_value' (621 MB). DHI recommends implementing regular maintenance tasks to wipe stored events and the change log, and utilizing a Time Series Provider to reduce the need for local copies of time series data.
• DHI identified several technical weaknesses in the BWDSS scripts, including a lack of error handling (try-except blocks), missing progress logging, and the use of hardcoded file paths. To improve resilience and maintainability, DHI recommends separating data from code by moving paths to spreadsheets or script arguments.
• A custom MIKE Workbench timeseries provider was developed to allow access to data stored in the 'dbo' PostgreSQL schema without duplicating data in the MIKE Workbench database. This provider uses specific stored procedures—fn_coct_mikeworkbench_timeseries_table, fn_coct_mikeworkbench_timeseries_values, fn_coct_mikeworkbench_single_value, and dbo.fn_coct_mikeworkbench_group_separator()—to retrieve time series properties and values.
• To integrate groundwater resources, the City of Cape Town plans to monitor approximately 400 groundwater boreholes in the Atlantis, Cape Flats, and Table Mountain aquifers. DHI recommends creating a new database schema specifically for these real-time groundwater observations to avoid overloading existing schemata.
• DHI provided a step-by-step guide for operationalizing transient FEFLOW groundwater models within the BWDSS. While two steady-state models for the Cape Flats and Steenbras Nardouw aquifers were tested and registered successfully, DHI noted that transient models are required for the intended purpose of simulating and forecasting groundwater levels to satisfy city water demand.

APA

WWF South Africa (2022). Technical Learning Exchange on groundwater data in decision support tools. https://wwfafrica.awsassets.panda.org/downloads/5_annexure_5_learning_exchange_partnering_with_dhi.pdf

Chicago

WWF South Africa. Technical Learning Exchange on groundwater data in decision support tools. 2022. https://wwfafrica.awsassets.panda.org/downloads/5_annexure_5_learning_exchange_partnering_with_dhi.pdf.

Wikipedia

{{cite report |author=WWF South Africa |title=Technical Learning Exchange on groundwater data in decision support tools |date=4 May 2022 |url=https://wwfafrica.awsassets.panda.org/downloads/5_annexure_5_learning_exchange_partnering_with_dhi.pdf |access-date=17 August 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{wwfsouthafrica2022technical, author = {{WWF South Africa}}, title = {{Technical Learning Exchange on groundwater data in decision support tools}}, institution = {WWF South Africa}, year = {2022}, month = may, url = {https://wwfafrica.awsassets.panda.org/downloads/5_annexure_5_learning_exchange_partnering_with_dhi.pdf}, urldate = {2026-08-17}, note = {Indexed by Climate Insights Directory} }