the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 27 Jul 2026
Large-scale atmospheric drivers of recent hydroclimatic changes in the Horn of Africa: evidence from the Awash Basin terminal lake system, Ethiopia
Abstract. The Horn of Africa, and specifically Ethiopia's Awash Basin, is experiencing accelerating hydroclimatic change, yet the mechanisms driving multi-decadal shifts in its regional water balance remain poorly understood. The terminal Afambo–Gemeri–Abhe lake system, an endorheic water body at the outlet of the Awash Basin, acts as a natural integrator of its catchment's net water balance, providing a sensitive record of regional atmospheric dynamics. This study reconstructs its hydrological history over 1985–2024 by coupling the GR2M hydrological model with a lake water-balance model, constrained by satellite observations and in situ data. The model reproduces observed lake surface area dynamics and reliably extends the record into the data-scarce pre-1998 period allowing us to relate recent lake surface area changes to large scale circulation variability and trends.
The reconstructed history reveals two superimposed signals. The first is pronounced variability across interannual to interdecadal timescales, driven by the El Niño–Southern Oscillation (ENSO), the Pacific Decadal Oscillation (PDO), and the Subtropical Indian Ocean Dipole (SIOD). The second is a non-oscillatory, multi-decadal expansion: the lake tripled in surface area from ∼150 km2 in the 1980s to ∼450 km2 by 2024, with growth accelerating markedly after 2020. This systematic expansion reflects a reorganization of the regional water balance driven by increased July–September rainfall and changes in potential evapotranspiration. We attribute these basin-scale changes to large-scale atmospheric circulation shifts, including a strengthening of the Somali Jet, the Tropical Easterly Jet, and the Indian monsoon circulation, and a northward shift of the Intertropical Convergence Zone and of the African Easterly Jet, likely linked to amplified regional warming over the Arabian Peninsula and the Tibetan Plateau. This modelling framework provides a transferable tool for evaluating past and future hydroclimatic variability, supporting more robust water resource assessments in data-scarce regions.
- Preprint
(18098 KB) - Metadata XML
-
Supplement
(3312 KB) - BibTeX
- EndNote
Status: open (until 15 Oct 2026)
- RC1: 'Comment on egusphere-2026-3724', Anonymous Referee #1, 24 Aug 2026 reply
-
RC2: 'Comment on egusphere-2026-3724', Anonymous Referee #2, 26 Aug 2026
reply
Review of “Large-scale atmospheric drivers of recent hydroclimatic changes in the Horn of Africa: evidence from the Awash Basin terminal lake system, Ethiopia” by Sierra et al.
General comments
The manuscript “Large-scale atmospheric drivers of recent hydroclimatic changes in the Horn of Africa: evidence from the Awash Basin terminal lake system, Ethiopia” by Sierra et al., addresses a timely and relevant topic concerning hydroclimatic variability and lake-level changes in the Awash Basin. One of its main strengths is the combination of multiple datasets, methodological approaches, and perspectives, which allows the system to be examined from several complementary angles. I found this broad approach potentially valuable and think the study could provide useful insights into the climatic and hydrological controls on recent changes in the basin.
At present, however, I have several substantial concerns regarding both the methodological basis of the analysis and the way the results are presented. Most importantly, the hydrological model plays a central role in the paper and underpins several of its main conclusions, yet its formulation, calibration, validation, parameter selection, sensitivity, and performance are described only very briefly. This makes it difficult to assess how much confidence can be placed in the model results, particularly given the apparent change in model performance around the large lake-level rise in 2020. Related assumptions concerning evaporation, groundwater and transmission losses, precipitation biases, and the treatment of the atmospheric water budget also require stronger justification and, where possible, explicit evaluation of their associated uncertainties and residuals.
More generally, the manuscript currently contains a very large number of datasets, analyses, and relatively short results subsections. While this diversity is one of the study's strengths, it also makes the main scientific narrative difficult to follow. In several places, results are presented only briefly before the text moves directly into interpretation and comparison with previous studies, making it difficult to distinguish clearly between what the analyses show and how the authors interpret those findings. This is particularly noticeable in the later parts of Sect. 3 and across Figs. 8–14. I therefore encourage the authors to revise the presentation: more clearly distinguish results from discussion even if the two remain within a combined section, and consider moving some of the supporting analyses and figures to the Supplementary Material. Doing so would help bring the central findings and the evidence supporting them much more clearly to the foreground.
Overall, I think the manuscript makes a very interesting contribution, but substantial revision is needed before the robustness of its main conclusions can be properly assessed.
Major comments
- 2.2.2: This section lacks details, and assumes both that readers are familiar with the model, and that Figure 2 is enough to trust the model. I disagree with these two: Can you elaborate on the model? What are the two parameters? How were they chosen? Was the model calibrated (besides the factor 2 for the high flows)? Validated? How was the factor 2 for the high flows decided? Were there any sensitivity tests for the model parameters and for the correction of the high flows? How were the other precipitation datasets used in the model? What are the metrics used for the calibration and validation and how do you define a success of the model? What are the calibration objectives (match total volumes, high flows, specific peaks, etc.)? How is the model performance over the gauges for which it was not calibrated? Does this matter for your study? L177: What do you mean by “bias-correction”? How come there is still a positive bias after it?
- L192–197: Can you think of a reason why GLEAM should perform OK for the other lakes but not for Lake Abhe? Are there other PET estimates for the region in contrast with GLEAM? In that case, why should we trust it in other places? Can the model mismatch come from a different factor rather than evaporation only?
- Groundwater: This is mentioned as a variable in Eq. 1 but later said to be neglected. Can you reinforce the fact it is negligible? How about the sandy river bottoms in the lower basin next to Lake Abhe? This factor can be super important in arid, sandy regions (Knighton and Nanson, 1994; Morin et al., 2009).
- 3.1: There seems to be a very big discrepancy between the model performance before the big 2020 rise and after it. This odd behaviour needs to be clarified by both describing the model validation process, and finding the reasons for this discrepancy.
- One of the strengths of this manuscript is the combination of multiple datasets, methodological approaches, and perspectives, which allows the problem to be examined from several complementary angles. At the same time, however, the large number of datasets, methods, and relatively short results subsections makes the overall flow of the paper difficult to follow. In particular, Figs. 8–12 are covered within only a few sentences, and it is often unclear where the presentation of results ends and the discussion begins. Please make a clearer distinction between the two, even if Results and Discussion remain combined. The reader should be able to distinguish clearly between what was found and the authors’ interpretation of what these findings mean, their possible causes, and how they relate to previous studies. I also suggest considering whether all of Figs. 8–14 are needed in the main text, particularly as many of their main messages are ultimately synthesised in Fig. 14. Some of these figures could perhaps be moved to the Supplementary Material/Appendix, while relevant background material from the opening paragraphs of these subsections could instead be incorporated into the Introduction. This is particularly relevant for Sect. 3.3, where the results themselves are described only very briefly before the text moves into an extended discussion. It may therefore be clearer to treat much of this section explicitly as Discussion, while summarising the key results more concisely and presenting only the main findings in one or two figures.
Specific comments
L26: “have” should be “may have / may impose” or be supported by a reference.
L32: What do you mean by “recurrent droughts”? That both the short and long rains are low, or that there are two consecutive drought years?
L33: 80% of the annual mean – is this for the entire country or the inner parts? These numbers are in contrast to those shown by Palmer et al. (2023). I think the clarification is done in L64, where the same numbers are stated for only the Awash basin. Please reconcile these two.
L97: One of the things that Dinku et al. (2018) mention is the bias CHIRPS data have over dry (and in some cases mountainous) regions. Moreover, Dinku et al. (2011) specifically describe the problem of below cloud rainfall evaporation in dry regions. Finally, GLEAM uses MSWEP precipitation as input. This means that by using other precipitation datasets you also inherit some error in the ET coming from the precipitation product used in GLEAM. Please address these issues.
L99: “TRMM 3B42 V7” – do you mean IMERG V7?
L118–120: Are there temperature discrepancies when using CRU data for mean T and ERA5 for T? Why do separate these into two datasets?
Sect. 2.1.7: Please explain: (1) How the Landsat-based extension (2021–2025) works; (2) what’s the reliability of the data from Zhao et al. (2022), which are based on GSWD, to the region; (3) what is the period of the HydroWeb span, and how reliable it is. If it does not, what did you do for the previous period? (4) what HydroWeb product was used?
Sect. 2.2.1: This section seems to resolve one of the problems solved by Zhao et al. (2022), which is one of the input sources. Isn’t it? Please explain.
Appendix: Do a and b correspond to x and y in the A1–A5 eqs.? Please make this clearer. Also, these equations assume no transmission losses. Can this be justified? How about irrigation abstractions along the rivers?
Fig. S4 and hypsometric curves: The figure is not clear enough – what are the Altimetry radar contours? Can you state their elevations? Elevation contours seem unphysical; consider either smoothing them, or using a different approach, e.g., the one from Armon et al. (2020). Can you validate the hypsometry using external data (e.g., ICESat-2)?
Sect. 2.3.2: Why don’t you use GLEAM to assess ET here? Is this consistent with your use in the hydrological model?
Why not using surface pressure to top of the model (or if it’s easier, 200 hPa)? I’m asking because in large parts of the region the surface is well above the 1000 hPa line. Do you take the lowermost layer or what? Additionally, given that the datasets are not equal, I would not expect a water balance closure. Can you quantify the residual using the different datasets?
Sect. 2.3.3: Given your precipitation-based ITCZ identification, please consider replacing the term ITCZ with a more appropriate term. Please see discussions in: Nicholson (2009; 2018).
L263–264: How can you tell this is the driver? If that’s the case, why isn’t it the same across the other parts of the time series?
L280–282: I’m not sure I can see this. Excluding the period of 1985–1990, they seem to be roughly static. Can you explain how you got to this conclusion? The same holds for the conclusion in 285–287. Please quantify/explain.
L289–291: Are these trends similar to gauge-based trends? How about other dataset? This question rises whenever a long-term satellite product is used, because sensors and satellites are being replaced, and it is hard to assess the validity of these trends.
L299: Can you explain what is this warming? Isn’t it also the human-induced warming?
L316–327: I suggest avoiding the interpretation of monthly mean 500-hPa vertical velocity as a direct proxy for convective activity. At ERA5 resolution, and particularly after monthly averaging, w500 primarily represents large-scale mean ascent/subsidence rather than convective updraughts. Indeed, the stated motivation for monthly averaging: to suppress short-lived convective variability, appears inconsistent with using the resulting field to characterise convection itself. I suggest describing this variable as an indicator of large-scale vertical motion or the large-scale dynamical environment associated with convection, unless its relationship with an independent measure of convective activity is demonstrated.
Fig. 6: Please add the basin shape to this figure. With regards to Fig. 6b, I think detrending is necessary, as we expect trends (in both datasets) to have an impact on this result.
Fig. 7: Why Lake Abhe only? Why not showing the observed lake time series too?
L356: How do you define the “boreal summer” is it the same as your long-rains definition?
L412: What is the meaning of “enhancing kinetic energy”? Shouldn’t the interaction with topography reduce the speed and kinetic energy of the flow?
Fig. 12e: Can you please add the 0-precipitation trend line? If all values are positive, you could put it in the lowermost part of the axis, but otherwise it’s hard to distinguish negative from positive trends.
L434: How can you tell that this increase in specific humidity comes from increased orographic precipitation? Can it also be there because of advection?
Sect. 3.4.2: The first two paragraph seem to be related to Sect. 3.4.1. Can you please put them together?
All figures presenting maps from -25 to 25: Consider reducing the latitudinal span of these figures. Perhaps 40W to 90E or something like this.
Fig. 13: Can you add the Sahara and Sahel rectangles related to Fig. 13b to one of the maps in this figure? Can you add political borders to panel c and d or zoom out a bit so its easier to read these maps? And the river catchment as well?
L484: I’m not sure I understand what the tool is. Is the problem stated common throughout the globe for pre-1998 data?
Technical corrections
L19: “region” should probably be “regions”.
Supplementary Material references are all corrupted. Please correct this.
L354: “Sect. X” should be corrected.
Fig. 8: Please make sure all colour axes are symmetric with respect to 0. Please make sure to do the same for all the corresponding figures.
L377: “how long term trends” seems to be missing “the” or “these”.
Fig. 9c: Are you sure about the *10^3 units? Should it be 10^-3?
L435: Missing space after “12e).”
L456: Missing reference (“?”).
References
Armon, M., Dente, E., Shmilovitz, Y., Mushkin, A., Cohen, T. J., Morin, E., and Enzel, Y.: Determining Bathymetry of Shallow and Ephemeral Desert Lakes Using Satellite Imagery and Altimetry, Geophys. Res. Lett., 47, 1–9, https://doi.org/10.1029/2020GL087367, 2020.
Dinku, T., Ceccato, P., and Connor, S. J.: Challenges of satellite rainfall estimation over mountainous and arid parts of east africa, Int. J. Remote Sens., 32, 5965–5979, https://doi.org/10.1080/01431161.2010.499381, 2011.
Dinku T, Funk C, Peterson P, et al. Validation of the CHIRPS satellite rainfall estimates over eastern Africa. Q J R Meteorol Soc. 2018; 144 (Suppl. 1): 292–312. https://doi.org/10.1002/qj.3244
Knighton, A. D. and Nanson, G. C.: Flow transmission along an arid zone anastomosing river, cooper creek, australia, Hydrol. Process., 8, 137–154, https://doi.org/10.1002/hyp.3360080205, 1994.
Morin, E., Grodek, T., Dahan, O., Benito, G., Kulls, C., Jacoby, Y., Langenhove, G. Van, Seely, M., and Enzel, Y.: Flood routing and alluvial aquifer recharge along the ephemeral arid Kuiseb River, Namibia, J. Hydrol., 368, 262–275, https://doi.org/10.1016/j.jhydrol.2009.02.015, 2009.
Nicholson, S. E.: The ITCZ and the seasonal cycle over equatorial Africa, Bull. Am. Meteorol. Soc., 99, 337–348, https://doi.org/10.1175/BAMS-D-16-0287.1, 2018.
Nicholson, S. E.: A revised picture of the structure of the “monsoon” and land ITCZ over West Africa, Clim. Dyn., 32, 1155–1171, https://doi.org/10.1007/s00382-008-0514-3, 2009.
Palmer, P.I., Wainwright, C.M., Dong, B. et al. Drivers and impacts of Eastern African rainfall variability. Nat Rev Earth Environ 4, 254–270 (2023). https://doi.org/10.1038/s43017-023-00397-x
Citation: https://doi.org/10.5194/egusphere-2026-3724-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 107 | 57 | 13 | 177 | 40 | 13 | 16 |
- HTML: 107
- PDF: 57
- XML: 13
- Total: 177
- Supplement: 40
- BibTeX: 13
- EndNote: 16
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The paper presents an interesting work, but needs some guidance to improve the presentation. There is too much going on, so a need to consolidate. This reviewer will focus on the results via the figures in sequence. Fig 1b,c could be deleted, Fig 2a,b the annual cycle 4m3/s and time series 1m3/s don't match, Fig 3 lake surface spread is interesting, but what about altimeter height, presumably corresponding? Fig 4 suggests that lakes are useful indicators of trends, rain and evap as oscillators, in (f) the pot.evap is dispersed, is there any way to say which is better? Fig 5 fine, Fig 6 clearly PDO influence, not ENSO - IOD, Fig 7 suggests the lake time series has 2 degrees of freedom, eg. one cycle, Fig 8 onward - delete climatology, too much information is diluting the message, Fig 9 wind arrows are too close horizontally, need to spread (sub-sample horizontally) and magnify vertical motion, shouldn't a Walker cell be analyzed E-W (not tilted?), and extend the analysis further east to Indian Ocean, Fig 10 trend of 200 wind useful, the rest are ambiguous and could be deleted, Fig 11 seems off-the-point, Fig 12 N-S section should extend northward to catch the Red Sea airflow, and wind arrows need to spread as noted above, Fig 13 fine, Fig 14 fine, but mid-troposphere and diabatic heating occur much higher than 700, more like 400-500 hPa.
With these changes in the results, naturally the supporting text can be focused and distractions can be eliminated to have a clear message. A useful reference might be: doi.org/10.3390/cli13100215