Sub-catchment vs Rain on Grid. Why should we care

The accurate representation of rainfall-runoff processes in urban environments is crucial for effective stormwater management and flood mitigation. Two primary approaches dominate the profession: Rain-on-Grid (ROG) and sub-catchment (lumped conceptual) models. Each possesses distinct strengths and weaknesses, impacting their suitability for specific applications.

Stantec, working with Wellington Water, recently updated an integrated 1D-2D urban stormwater model of the Wellington CBD from a sub-catchment to a ROG model as part of Let's Get Wellington Moving -Golden Mile (GM). This transition was completed to quickly update the model and make changes to the surface without having to delineate subcatchments for each variation of the design. It also allowed for the simulation of water on the surface providing insight into the placement of sumps, kerbs and raised pedestrian crossings. As part of the update, Stantec compared baseline ROG results against baseline subcatchment results at multiple locations in both 1D and 2D domains. The results showed minimal changes in flow rates and overall flood extents in and around the GM corridor.

However, there were some observed locations where there were significant changes to flow rates in both the 1D and 2D domains. These changes were primarily driven by small misalignments in the GIS network and ground surface. This was more prominent in the steeper hilly areas where overland flow paths were narrower and more constrained. This led to lower flow in the pipes and more flow on the surface. Despite these changes, the modelled results show the changes were not enough to impact GM as the proposed corridor is along flatter land close to the coast, and thus the changes are dwarfed by the sheer volume of water already on the surface from other much larger effects such as sea level rise.

Despite it not being an issue for GM, the misalignment of the GIS data and Lidar can be an important aspect to understand as it can significantly alter the flow rates depending on the conditions. Conversely this “error” may not be an error. Countless videos including recent observed flooding in Wellington, show that fast-moving water can flow straight past or over sumps with minimal water entering the chamber due to the high velocity and or blockages from debris. The sub-catchment method, although less prone to the misalignment of the GIS and LiDAR, assumes all water within the delineated catchment reaches and enters the sump. Potentially missing out on key overland flow paths.

Overall, each method has its positives and negatives, and like many other modelling aspects, it is heavily reliant on the input data. Thus, it is important to know the limitations of the modelling approach and the data that you are working with.