How (Managed) Drainage Shapes Regional Hydrology: Model Based Insights for Flanders
Abstract. Artificial drainage is widely used in agricultural landscapes to improve trafficability and prevent waterlogging, but comes at the expense of groundwater storage and regional drought resilience. Through model scenarios, this study evaluates how large-scale drain-level adjustments influence groundwater dynamics, water balances, and streamflow extremes across different regions in Flanders, Belgium. A physically based, spatially distributed hydrological model coupling land-surface and groundwater processes was adapted to represent controlled drainage through adjustable weirs, by modifying drain levels and incorporating temporary water storage within drainage ditches behind control structures. Permanently raising drain levels from 1 m to 0.5 m below the surface increased average groundwater levels by 15 cm across Flanders, with increases exceeding 30 cm in actively drained areas. Higher groundwater availability generally resulted in modest increases in evapotranspiration, particularly in flat sandy regions. Simulations showed that raising drain levels shortly after winter (April) was more effective than implementing measures only during summer (June onward), when groundwater tables had often already fallen below the drainage depth. Drainage management also altered streamflow extremes. Under a permanent drain-level increase, low flows increased by 12.8 %. In contrast, summer-only management reduced low flows by up to 10.1 % in some drainage-dominated catchments, highlighting a trade-off between local water retention and downstream water availability. The effects on flood peaks were spatially variable. Peak discharges increased by up to 11.3 % in saturation-prone areas due to reduced unsaturated soil storage, whereas temporary storage behind control structures reduced peak flows by up to 26.5 % in regions dominated by infiltration-excess runoff. The results demonstrate that drainage management can be an effective tool for increasing groundwater storage and mitigating drought impact. Nevertheless, its impact on flood risk and downstream flow regimes needs to be assessed carefully, since it depends strongly on local hydrogeological conditions and management timing. Depending on the location, the effect can exacerbate or mitigate flooding risk due to peak flows. These findings highlight the need for spatially targeted and adaptive drainage-management strategies to support climate-resilient water management in heavily modified agricultural landscapes.
General: the authors provide a regional modeling analysis on the impact of controlled drainage (both tile drains and ditches) on regional water retention, in relation to crop water availability and effect on several hydrological fluxes. The study is well embedded in climate adaptation strategies for Flanders. Although the paper is well structured and clearly written, it could be improved in clarity and impact at several points.
The main aspect to clarify is the definition of controlled drainage. This is often used only for the adaptation of tile drainage systems, but the paper especially deals with weirs in the surface water network. This needs to be clarified early in the paper. Additionally, it is not clear how sites with tile drainage are included. Maybe they are not included, and the research only considers ditches + weirs?
It should also be described more clearly how the surface water levels are modeled and how controlled drainage is implemented in the model. It should be implemented as weir levels that are adjusted from 1m-ss to 0.5m-ss, and ditch water levels should be modeled dynamically. However, it is not clear how dynamic surface water levels are modelled. But maybe the authors didn’t use dynamic modeling of surface water levels (and herewith the drainage base), but used fixed surface water levels of 1m-ss and 0.5m-ss. If this is done, the impact of raising the drainage base on simulated gw levels is way too positive, as in practice, these surface water levels cannot be maintained without external water supplies (e.g. from larger river systems). This is a crucial aspect which needs to be clarified and should be clearly described in the methodology section. I would also recommend to include figures supporting the used modeling concepts, especially on GW-SW interactions, and modeling SW dynamics, as they are key for the goals of the research. Related to this: L86: provide analysis that shows that the scale of 250m indeed is suitable for the purpose of the research.
The study aims to analyze impact on growing conditions for crops, with concepts from Aquacrop to simulate plant water stress as a function of too dry or too wet conditions. More process-based approaches are available though and choices in modeling concepts can significantly impact modeling results (see e.g. Van den Brink et al. 2026 Effect of increasing conceptual model detail on simulated crop yield-drainage base relations). The authors are encouraged to include a discussion on this, as analyzing the impact of soil moisture and soil oxygen conditions on crop growth is among the key components of the study.
Regarding oxygen / waterlogging stress specifically: authors state that they model aeration stress in areas that are persistently wet (both section 3 and 4). It is questionable if agriculture would be feasible anyway in these areas, and if so, this would likely be grasslands, of which the species probably have been adapted to these persistent wet conditions (e.g. by creating shallow rooting systems). Doing so, they limit oxygen stress. The authors model oxygen stress in wet areas, because such crop adaptations are not incorporated in the modeling approach. A discussion is needed on this.
The results are tailored to Flanders, but the broader scientific community could only benefit if the results and insights are made more general. This should be possible. Also include a description for regions for which the results and methods are most relevant. I guess it is limited to regions with a yearly precipitation surplus (temperate climates). What can people learn from your insights? Where to put more effort on water retention and where as it limited impact on freshwater availability? And why? So, what could other regions learn from these (generally applicable) insights?
L105: abstractions: do these include agricultural abstractions? And if so, what is the effect of using constant daily abstractions?
L134-135: model evaluation for hydrographs. However, in the results evaluation is also done for GWL. Please include the use of GWL for evaluation in the methods section too.
Section 2.2: please improve the description of how controlled drainage is modelled, as adjustment of the weir level (ditches) or threshold level in control pits for controlled tile drainage, is not the same as adjustment of the surface water level / drainage base. This should be clarified as it is a critical point for the study as a whole (my interpretation from the results is that surface water levels are modeled dynamically, and only weir levels are adjusted, but please include a clear description in the methods section). Also reconsider the use of terminology (drain level, vs weir level vs sw water level vs control pit water level vs …) and be very consistent throughout the paper.
To continue on this aspect: modeling surface water level dynamics could be different in free draining areas vs. polder areas and areas with/without water supply from external sources to maintain specific sw levels. Would be good to include this in the revised description.
And: Infiltration by ditches: generally only possible if SW>GWL, which often requires external water supply to maintain SW levels.
L180-181: unclear sentence, please rephrase
L227 remove bracket ‘Fig 3)’
L242: poor model performance with deeper GWL: Ok, but deeper gw levels might also be less relevant for crop water availability, so is this a problem?
Fig 4: correlatie -> correlation.
Additionally, what correlation is used? Pearson, Spearman? And only R or R^2. If R (and I think you did) than R^2 for GWL relative to ss has a very poor correlation. You might even say that the suitability for modeling GWL is too weak? Please elaborate on this aspect, including a discussion about the analysis the model can and cannot be used for. I can imagine that for detailed analysis between the impact of raising the drainage levels and crop yields, a fields scale agrohydrological model might be more suitable.
Fig 6: A: waterbalance -> water balance; Evaporation = ET? And does ET include interception evaporation? The scales in the legend are not so clear (are the number the highest and lowest values?); GW Recharge: only to phreatic GW, right, and not recharge of deeper aquifers. Recharge of phreatic GW might be less relevant if this water is quickly discharged to SW. Please include a short discussion.
L344-345: include in methods section
Section 5 includes potential improvements to be made. However, this study could profit by including (some) more detailed analysis of GW-SW interactions. Don’t just note that this is possible, but already introduce some insights in the key processes and how ‘process-based’ they are actually incorporated (also see previous comments), and what this means for the interpretation and general applicability of the results. Include a discussion on the regions for which the results are relevant.
Â
specific questions:
1.      Does the paper address relevant scientific questions within the scope of HESS? Yes
2.      Does the paper present novel concepts, ideas, tools, or data? Yes
3.      Are substantial conclusions reached? Should be improved
4.      Are the scientific methods and assumptions valid and clearly outlined? Needs improvement
5.      Are the results sufficient to support the interpretations and conclusions? Yes but needs improvement
6.      Is the description of experiments and calculations sufficiently complete and precise to allow their reproduction by fellow scientists (traceability of results)? Some improvement needed
7.      Do the authors give proper credit to related work and clearly indicate their own new/original contribution? yes
8.      Does the title clearly reflect the contents of the paper? Yes
9.      Does the abstract provide a concise and complete summary? Yes
10.  Is the overall presentation well structured and clear? Yes
11.  Is the language fluent and precise? yes
12.  Are mathematical formulae, symbols, abbreviations, and units correctly defined and used? Yes
13.  Should any parts of the paper (text, formulae, figures, tables) be clarified, reduced, combined, or eliminated? no
14.  Are the number and quality of references appropriate? yes
15.  Is the amount and quality of supplementary material appropriate? yes
Â
Â