Preprints
https://doi.org/10.5194/hess-2023-68
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/hess-2023-68
© Author(s) 2023. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
27 Mar 2023
Status: this preprint is currently under review for the journal HESS.
On the optimal level of complexity for the representation of wetland systems in land surface models
Mennatullah Tarek Elrashidy
CORRESPONDING AUTHOR
Department of Civil, Environmental and Geological Engineering, University of Saskatchewan, Saskatoon, Canada
Andrew Ireson
Department of Civil, Environmental and Geological Engineering, University of Saskatchewan, Saskatoon, Canada
School of Environment and Sustainability, University of Saskatchewan, Saskatoon, Canada
Saman Razavi
Department of Civil, Environmental and Geological Engineering, University of Saskatchewan, Saskatoon, Canada
School of Environment and Sustainability, University of Saskatchewan, Saskatoon, Canada
Abstract. Wetland systems are among the largest stores of carbon on the planet, most biologically diverse of all ecosystems, and dominant controls of the hydrologic cycle. However, their representation in land surface models (LSMs), which are the terrestrial lower boundary of Earth system models (ESMs) that inform climate actions, is limited. Here, we explore different possible parametrizations to represent wetland-groundwater-upland interactions with varying levels of system and computational complexity. We perform a series of numerical experiments that are informed by field observations from wetlands in the well-instrumented White Gull Creek in Saskatchewan, in the boreal region of North America. We show that the typical representation of wetlands in LSMs, which ignores interactions with groundwater and uplands, can be inadequate. We show that the optimal level of model complexity depends on the land cover, soil type, and the ultimate modelling purpose, being nowcasting and prediction, scenario analysis, or diagnostic learning.
How to cite. Elrashidy, M. T., Ireson, A., and Razavi, S.: On the optimal level of complexity for the representation of wetland systems in land surface models, Hydrol. Earth Syst. Sci. Discuss. [preprint], https://doi.org/10.5194/hess-2023-68, in review, 2023.
Download & links
Mennatullah Tarek Elrashidy et al.
Status: final response (author comments only)
Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor
| : Report abuse
-
RC1: 'Comment on hess-2023-68', Anonymous Referee #1, 01 May 2023
General Comments
Here the authors explore different model complexities and configurations to highlight the need for coupling upland-wetland interactions in land surface models to better capture downstream hydrologic fluxes. In general, the inclusion of wetlands in LSMs is an important consideration for many types of landscapes. Overall, the manuscript was clear and the approach was sound. Below are some points that I think the authors should consider to help clarify some points for the reader.
One broad comment is that there is a bit of a disconnect in how the manuscript was framed (the title, abstract, introduction had a heavier emphasis on wetlands in general) vs the actual analysis and discussion (much larger emphasis on fens, which is the study site). The conclusions were logical based on the data from the fen, but it would be helpful to a) emphasize that fens by definition receive significant amounts of groundwater inputs and b) how other types of wetlands with different levels of GW connectivity could change your conclusions.
Specific Comments
- It would be helpful to reference more existing literature on the need for including proper interaction between wetlands and the upland to help strengthen the case for this particular study. While there may not be as many LSM studies that look at this directly, drawing parallels to watershed scale studies, which has quite a few studies in the recent years.
- Minor suggestions for Figure 1
- Expand part d to be larger to match the other panels for clarity;
- add the location of POJP piezometer for clearer connection to panel c);
- if possible, indicate the extent of the fen in part d) just to give some context for the reader. I understand that this can be variable throughout the fen. If the fen goes beyond the transect, ignore this comment!
- L123: The statement regarding the amount of surface water vs groundwater into the fen seems something unique to this system, and isn’t necessarily a feature of the V2 configuration – consider putting this elsewhere. As a side note, because I don’t have much context to the wetland:upland ratio or the water balance, the reader might be surprised by this statement. Might be worth indicating the relative areas for the wetland + upland and/or some estimated water balance in the site description
- L137: does the MESH-CLASS model have saturation-excess and infiltration-excess runoff components to determine runoff vs recharge? Similarly, how does it calculate the R flux? I would not expect a full description of the model, but since this is a major connection to the GW component, it would be helpful to briefly describe it. Also, is snow accumulation and melt modelled in the upland? I assume so, but since it’s explicitly mentioned in the fen model, but not here, it can create confusion
- L148: I may be mistaken, but I don’t think Figure 1 show the groundwater divide/no-flow boundary condition (this is in Fig 2?)
- L151 and paragraph: would be helpful to mention how dx is determined within the Darcy’s Law calculation
- L165: Units are not consistent across equation. If Ro is upland runoff (m3/d) and L is the hillslope length (m), the units are [L4/T]; similarly the (Rg+M-Ef)wf component have units of [L2/T], while the Q terms are [L3/T].
- L168: Are Cspill, hspill, n calibrated parameters, and is the model sensitive to them? I believe they are not referenced again later on but one might assume that they have major roles in changing Q.
- L181: Is there any downstream gage to calibrate? I would be hesitant to say that calibrating the GWT in the upland represents the performance of the collective fen-upland-GW models - especially since three of them do not have the backwards interaction in the model structure
- L185: were these all generated from uniform distributions?
- L186: maybe use “(threshold is chosen arbitrarily based on…)” instead of “chosen rather arbitrary”
- L189: I would consider putting in the best parameter set in Table 1 to give context for readers rather than just embedded in the text later on
- L194: are these L values corresponding to likely wetland-upland areas?
- Figure 3: Should write in caption the simulation number (V1?)
- Figure 6d: missing y axis label to be consistent
- L291: It is slightly hard to understand why at low L values, the groundwater table does not fluctuate in the chained model – it would be worth discussing why. I would assume there’s still stochastic inputs to the groundwater from recharge/precipitation, and it’s not that the groundwater table has reached the lower boundary/bedrock
- L306: Land is capitalized mid-sentence
- L306: Because the reader does not know how the forest and grass (is it solely a runoff-coefficient difference, or does ET get affected too?) affects the model fluxes, it’s hard to attribute the changes in the model to solely the soil properties, which is the focus of this section.
- L313: I think that for this instance, it is true that the chained approach is adequate to illustrate the coarse grained soil texture. But I think it’s worth commenting that in areas of smaller hillslopes/contributing areas, that may not be the case (as proven in your previous experiment)
- L346: I would not necessarily include wetlands in ‘fen/wetlands’ as fens by definition have a lot of GW inputs. Having wetlands here can cause readers to assume that wetlands that either have more bi-directional interactions with the upland via groundwater, or don’t receive groundwater, should be treated the same way. While the authors wouldn’t run more simulations to capture other wetland types, I think it’s a valuable discussion point
Citation: https://doi.org/10.5194/hess-2023-68-RC1 -
RC2: 'Comment on hess-2023-68', Anonymous Referee #2, 30 May 2023
# SummaryThis paper explores how to adequately represent wetlands in land surface models, with a case study in White Gull Crrek, Saskatchewan. The authors find that existing parameterizations which ignore groundwater and upland influences to be inadequate and make some recommendations for how to improve the representation of wetlands in LSMs. The authors explore four different model representations, from fully uncoupled to fully coupled. The main conclusion of the paper is that explicit groundwater interactions must be accounted for the adequately represent wetlands in LSMs, with chained and fully coupled modeling approaches working well under different circumstances. Overall I think the paper was interesting, and straightforward to understand. However, I think this paper is only able to make a conceptual argument for this case, rather a quantitative one. The only comparison to observations comes from the comparison of the groundwater level in the upland site, which tests the ability of the model to represent the upland, but not the wetland. Of course, I am sympathetic to the fact that observations of discharge from the fen don't exist, making this a large challenge.However, I think the paper would be stronger if it could make a quantitative argument for the need for coupled models in representing the wetlands component, rather than a conceptual one. I think some additional analysis looking at either some downstream discharge could be useful for making this point, if such a stream gauge exists. I think correlating such a streamflow with the daily outflow into the river from your model setups could yield an interesting result. Similarly, correlating the $E_f$ from your fen model with latent heat from the FEN tower vs the OJP tower could make a good case for including the coupling between the wetland and upland. Because of these recommendations for additional analysis, I am recommending major revisions.# Major commentsAs of section 2, there is no real mention for the digression into coupling LSMs with GWMs as it relates to wetland modeling. Perhaps a few sentences connecting these topics and prior work on this connection would be useful.I was unclear where the information from the FEN tower actually comes in. At first it seemed like it was the data used to force the wetland model, but later it seemed like OJP was the site where the meteorologic data was taken. Could you clarify which datasets were used to drive which models?Equation 3 shows that there are several parameters, $c_{spill}$, $h_{spill}$, and $n$, whose values don't seem to be given anywhere. I am not sure how to interpret the results in section 5.2 without knowing these values. To me, it looks like the uncoupled model (V1) simply shows no outflow because $h_f$ is always less than $h_{spill}$, but I am not sure if this is correct. If so, doesn't that correspond to starting the simulation so that $h_{spill}$ is simply set to whatever the initial fen water level is? Wouldn't modifying the parameter values change the results? Even if you do modify these values, of course you will find that you just don't have enough water to maintain the fen's water level because precipitation is not enough to maintain the water level, but showing how these parameters affect this result would be useful and probably highlight your point that the coupling is necessary to maintain predict reailistic discharge from the fen.While you say the data is available, the code to reproduce the experiments does not seem to be. This should be linked in the "Data Availability" section.# Minor commentsThe term $L$ is never defined in equation 2.Line 186-7: "chosen rather arbitrary" -> "chosen arbitrarily"Line 238: You compare V1 and V3 and don't include V2 because it is functionally the same as V1 for the uplands, but having this spelled out here would here would be useful for anyone who skimmed over 5.1. Maybe just say (V1/V2)?Line 280: "Considering the GW dynamics underneath the upland is essential" Could you elaborate what it is essential for?Figure 6: Need to fix "(Error! Reference source not found.)" in the caption.Citation: https://doi.org/
10.5194/hess-2023-68-RC2
Mennatullah Tarek Elrashidy et al.
Mennatullah Tarek Elrashidy et al.
Viewed
Total article views: 446 (including HTML, PDF, and XML)
Cumulative views and downloads
(calculated since 27 Mar 2023)
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 325 | 113 | 8 | 446 | 2 | 4 |
- HTML: 325
- PDF: 113
- XML: 8
- Total: 446
- BibTeX: 2
- EndNote: 4
Viewed (geographical distribution)
Total article views: 435 (including HTML, PDF, and XML)
Thereof 435 with geography defined
and 0 with unknown origin.
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1
Latest update: 08 Jun 2023
Short summary
Wetlands are important ecosystems that store carbon and play a vital role in the water cycle. However, hydrological computer models do not always represent wetlands and their interaction with groundwater accurately. We tested different possible ways to include groundwater-wetlands interactions in these models. We found that the optimal method to include wetlands and groundwater in the models is reliant on the intended use of the models and the characteristics of the land and soil being studied.
Wetlands are important ecosystems that store carbon and play a vital role in the water cycle....