the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Jan 2026
Wildfire-induced disruptions to evapotranspiration, runoff, and water balance closure across California's water supply watersheds
Abstract. Wildfire activity has intensified across forested mountain watersheds globally, yet the basin-scale hydrologic consequences of large, high-severity fires remain poorly quantified. Here we integrate four decades of satellite-derived evapotranspiration (ET), precipitation (P), full natural flow (FNF) records, and spatially explicit fire-perimeter data to evaluate how wildfire alters ET, basin outflow, and water-balance closure across major water-supply basins in California. High-severity fires consistently suppressed ET by 100–250 mm in the first postfire year, with recovery strongly modulated by vegetation traits, moisture availability, and disturbance recurrence. Structurally diverse and moisture-rich basins recovered 75 % of prefire ET within 4–5 years, whereas drier, conifer-dominated systems required up to a decade. Although interannual P remained the dominant control on basin outflow, reduced ET partially offset drought-year declines in FNF within heavily burned sub-basins, indicating a localized compensatory effect. Water-balance analysis revealed systematic negative residuals (P − ET − FNF) during years with substantial fire disturbance, demonstrating measurable departures from steady-state closure. Basin-specific diagnostics showed that these deviations arise from both disturbance-driven hydrologic shifts and observational uncertainties, including precipitation underestimation and stream-gauge bias. Proportional and two-parameter adjustments improved closure across most basins, underscoring the need for disturbance-aware calibration in regional water-balance assessments. Collectively, our findings reveal that wildfires act as short-term hydrologic shocks that suppress ET, alter basin outflow patterns, and distort modeled water budgets across fire-prone headwater systems. Incorporating fire history, disturbance intensity, and ET-recovery patterns into hydrologic models and reservoir operations will be essential for improving postfire flow prediction and sustaining long-term water-supply reliability in an increasingly disturbance-affected climate.
- Preprint
(2086 KB) - Metadata XML
-
Supplement
(2394 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-6193', Anonymous Referee #1, 06 Feb 2026
-
AC1: 'Reply on RC1', Han Guo, 01 Mar 2026
We thank Referee #1 for the careful and constructive review. We agree that the points raised are important for strengthening the clarity and physical interpretation of the manuscript. These comments highlight aspects of the water-balance framework that benefit from more explicit explanation, and we will address them accordingly.
We agree that basin water storage (ΔS) is a fundamental component of the annual water balance and deserves clearer discussion. In the current manuscript, this aspect was not sufficiently articulated. Our intention was not to assume ΔS = 0 in a strict sense. Rather, P − ET was used as a flux-based comparison term against FNF to examine effective closure behavior at annual timescales. As the reviewer notes, deviations between P − ET and FNF may reflect a combination of interannual storage variability (ΔS), precipitation uncertainty, and streamflow measurement bias, particularly in basins with volcanic terrain or strong groundwater influence. We will clarify the interpretation of the P − ET − FNF residual as an effective closure term that may integrate storage variability and observational uncertainties, rather than as a strict error metric. We will also expand discussion of how post-fire shifts in infiltration, recharge, or groundwater contribution could influence this residual, especially in storage-sensitive basins.
We agree that the analytical thresholds (75% recovery benchmark, 3% burned-area threshold, and 500 m reference buffer) require clearer justification. These thresholds were selected as operational criteria to balance detectability, comparability, and sample robustness, but the rationale was not sufficiently documented. We will therefore:
- Provide supporting references for fractional recovery thresholds in post-fire studies. Recovery criteria based on 70–80% return toward pre-disturbance conditions have been applied in remote sensing analyses of post-fire vegetation and functional recovery (e.g., White et al., 2018; White et al., 2022). The 75% benchmark was adopted to represent substantial functional recovery while avoiding the unrealistic requirement of full (100%) return in semi-arid systems.
- Clarify the logic behind the 500 m buffer in terms of spatial comparability and reduction of edge effects. Similar buffer distances (on the order of several hundred meters) have been used in post-fire and remote sensing studies to limit spatial autocorrelation while preserving environmental similarity between disturbed and reference areas (e.g., Jin et al., 2012; Reddy et al., 2015; Dias & Acácio, 2024). We will explain how this distance balances independence and site similarity.
- Justify the 3% basin burned-area threshold within the range used in prior basin-scale hydrologic studies. Published thresholds have ranged from approximately 1% (e.g., Hallema et al.) to 5% (e.g., Beyene et al., 2022), reflecting trade-offs between disturbance detectability and statistical robustness. Our choice of 3% lies within this range and was intended to balance signal detectability with retention of sufficient fire-year observations. We will also include a brief sensitivity analysis using alternative threshold values to evaluate robustness.
We appreciate the reviewer’s concern regarding potential circularity within the CECS framework. Although CECS includes an internal water-balance structure, our closure analysis does not rely on CECS-modeled runoff. Precipitation is derived from PRISM (downscaled within CECS), ET is estimated independently using a remote-sensing and process-based modeling framework, and FNF is independently observed from CDEC. Thus, P, ET, and FNF are not internally constrained in the closure comparison. We will clarify this methodological structure more explicitly and discuss potential error propagation in a more transparent manner.
We agree that attributing closure deviations to precipitation bias requires stronger support. We will compare PRISM-derived basin precipitation with an independent gridded precipitation product over the overlapping period and present the results in the Supplement. This will allow a more evidence-based discussion of whether systematic precipitation differences align with the observed residual patterns.
We also agree that scale mismatch between localized burned areas and basin-integrated FNF stations is a key interpretive limitation. We will more clearly describe the spatial and hydrologic mismatch among (i) burned patches, (ii) basin-scale aggregation of ET and P, and (iii) downstream FNF measurements. We will emphasize that detectable runoff responses are expected primarily when a substantial fraction of the contributing area is burned or when disturbances overlap hydrologically sensitive zones. This clarification will help frame weaker runoff responses as consistent with precipitation dominance and basin storage dynamics at annual scales.
We appreciate the reviewer’s detailed and constructive comments. Addressing these points will strengthen the clarity and physical grounding of the water-balance framework. A structured, point-by-point response will be provided.
Citation: https://doi.org/10.5194/egusphere-2025-6193-AC1
-
AC1: 'Reply on RC1', Han Guo, 01 Mar 2026
-
RC2: 'Comment on egusphere-2025-6193', Anonymous Referee #2, 28 Feb 2026
Thank you for the opportunity to review “Wildfire-induced disruptions to evapotranspiration, runoff, and water-balance closure across California’s water supply watersheds”. In this paper the authors present a multi-decadal synthesis of wildfire, ET, precipitation, and full natural flow data across key California basins. The results are consistent with growing body of post-wildfire hydrologic response literature, making the topic timely and likely of interest to a broad audience.
I have several broader conceptual concerns that I believe should be addressed to strengthen the interpretation of the results. First, I am uncertain about the physical interpretation and motivation behind the parameter adjustments (scaling + intercept) applied to precipitation and/or FNF to improve water-balance closure. While these adjustments are presented as diagnostic tools, they risk appearing ad hoc without clearer linkage to quantified uncertainty or process understanding. I think it would strengthen the manuscript to clarify whether these adjustments are intended to identify systematic dataset biases, to approximate unaccounted storage fluxes, or to represent measurement error.
Second, the water-balance framework assumes that annual changes in storage (ΔS) are negligible when comparing P−ET to FNF. While this generally seems OK at the scales considered here, deviations between P−ET and FNF during fire years could plausibly reflect changes in storage rather than solely disturbance-induced hydrologic “imbalance” or observational uncertainty. I encourage the authors to more explicitly justify the ΔS ≈ 0 assumption, and discuss its limitations.
Finally, burn severity and vegetation controls are repeatedly invoked as key determinants of ET suppression and recovery, yet it is not fully clear how burn severity metrics are incorporated into the analysis. The methods primarily describe burned perimeters, and additional detail on whether severity (e.g., dNBR or similar indices) was used would improve clarity. Similarly, because vegetation structure and regrowth are central to much of the discussion of ET recovery trajectories, incorporating quantitative vegetation indices would help connect the mechanistic discussion to the results.
I’ve included some line-by-line comments below.
Abstract:
Line 26: Should “Interannual P” be “annual P”?
Lines 34-36:
Introduction:
L 49: What are the compounding disturbances? Repeated wildfires? Wildfire and drought?
L57: I’m not sure what “vegetation mortality surplus” means? Here and throughout it might be helpful to add a little more detail about the expected change to individual ET fluxes. The post-fire signal is most likely driven by decreasing transpiration, but there can be compensatory increases in ground evaporation. It might be worth referencing some of the papers by Collar et. al.
- Collar, N. M., Ebel, B. A., Saxe, S., Rust, A. J., & Hogue, T. S. (2023). Implications of fire-induced evapotranspiration shifts for recharge-runoff generation and vegetation conversion in the western United States. Journal of Hydrology, 621, 129646. https://doi.org/10.1016/j.jhydrol.2023.129646
- Collar, N. M., Saxe, S., Rust, A. J., & Hogue, T. S. (2021). A CONUS-scale study of wildfire and evapotranspiration: Spatial and temporal response and controlling factors. Journal of Hydrology, 603, 127162. https://doi.org/10.1016/j.jhydrol.2021.127162
L 80: and moisture availability? Especially in California, which is likely a moisture limited system?
L 88: While I generally agree that that this gap exists, there are quite a few papers in this space t(some of which the authors already reference) that do address this gap. For example:
- Goeking, Sara A.; Tarboton, David G. 2020. Forests and water yield: A synthesis of disturbance effects on streamflow and snowpack in western coniferous forests. Journal of Forestry. 2020: 172-192.
- Hallema, D.W., Sun, G., Caldwell, P.V. et al. Burned forests impact water supplies. Nat Commun 9, 1307 (2018). https://doi.org/10.1038/s41467-018-03735-6
- Beyene MT, Leibowitz SG, Pennino MJ. Parsing Weather Variability and Wildfire Effects on the Post-Fire Changes in Daily Stream Flows : A Quantile-Based Statistical Approach and its Application. Water Resour Res. 2021 Aug 31;57(10):1-20. PMID: 34898727; PMCID: PMC8654146.
L118 – Some recent work has identified aridity as an important driver of post-wildfire hydrologic response. Is there an aridity gradient across your watersheds? It might be worth including in Table 1
- Baudena M, Santana VM, Baeza MJ, Bautista S, Eppinga MB, Hemerik L, Garcia Mayor A, Rodriguez F, Valdecantos A, Vallejo VR, Vasques A, Rietkerk M. Increased aridity drives post-fire recovery of Mediterranean forests towards open shrublands. New Phytol. 2020 Feb;225(4):1500-1515. doi: 10.1111/nph.
Data section.
As a reader who is not familiar with these particular hydrometeorological datasets it would be helpful to include a litter more detail describing how they are calculated. For example from the spatial resolution I assume that the ET data is derived from an energy balance model built on Landsat imagery, but maybe that’s incorrect? Does the ET model make assumptions over burned areas that could influence the ET estimates? Is P derived from the same model? Does it incorporate any other precipitation datasets?
Similarly with the FNF data. Do these data account for diversions and reservoir operations? It might be helpful to include a little more detail since you reference reservoir effects a few times in the discussion.
L 160-161: Can you clarify that this is mean annual ET? And is the post-fire ET just a single year average?
L 186-196: I appreciate the discussion around uncertainty and challenges it could pose for interpretation. It seems like these are real challenges when considering the interacting effects of P and ET in a post-fire setting. I’m not quite sure I follow the motivation of your proportional adjustments approach though. Perhaps this an attempt to uncover systematic biases in the datasets? Is the assumption that d_S is really zero? Doesn’t a land cover disturbance like wildfire kind of inherently violate the assumption of steady state?
L 203 – 217: Interesting to see the range of recovery trajectories. Are there and vegetation indices that could help explain why some burned areas seem to take so long to recover? You mention vegetation structure as a control on ET recovery a few times in the discussion, but I didn’t see it really built into the results.
L 215: Here and elsewhere you reference burn severity, but it seems like your only considering burned perimeters? Is burn severity built into the analysis? If so, more details would be helpful
L 218-219: Maybe consider moving this sentence in the methods?
L239-240: Im not sure I understand the ET recovery at the watershed scale. Don’t the repeated fires in the basin (even if they aren’t spatially overlapping) continually move the “baseline”? There were likely fires prior to the start of your analysis as well. Over the larger basins, there might not really be such a thing a “prefile ET”?
L248: maybe consider replacing dry years with precipitation and temperature?
L 250: Could you describe what you mean by ET playing a small but detectable role in driving FNF? Aren’t interannual fluctuations in ET expected? Especially in a moisture limited system?
L 283: I know you mention this later, but a similar FNF exceedance seems common in the DAV basin. The sample size might be too small, but a basin by basin comparison might be interesting as well. I’m also curious how long this difference persists, is it significant only the first year following fire?
Figure 5:
Please add labels to the y axis.
L 300-301: Couldn’t changing baseflow/groundwater contribution be another explanation?
L 303: I understand the inclination to point to observational uncertainties, but couldn’t a change in storage also explain FNF < P-ET?
L 307: Again I wonder if considering change in storage is appropriate? Or maybe even lagged response?
L 308: The intercept parameter wasn’t mentioned in the methods section. Can you help the reader interpret the physical implications of this parameter?
Discussion:
L 327: How was burn severity incorporated?
L 336-347: This is where I think some further analysis of vegetation indices could help inform your recovery discussions. If they are in the SI it might be worth bringing them into the main ms.
L 339: Are you referring to individual burned areas within Pit or the basin total?
L 434: This was the first mention of reservoir operations. As I mentioned above, some explanation of how they are accounted for in the FNF data might be helpful
L 464: There are a few recent reviews of post-wildfire hydrologic modeling, it might be worth including a reference here.
Citation: https://doi.org/10.5194/egusphere-2025-6193-RC2 -
CC1: 'Reply on RC2', Han Guo, 01 Mar 2026
We thank Referee #2 for the careful and constructive review. We appreciate the thoughtful conceptual questions raised regarding the physical interpretation of the proportional adjustments, the role of annual storage variability (ΔS), and the incorporation of burn severity and vegetation controls.
The points concerning the diagnostic nature of the scaling and intercept parameters, and their linkage to systematic dataset bias versus storage variability, are particularly helpful. We agree that clearer framing is needed to avoid misinterpretation of these adjustments as ad hoc corrections rather than exploratory diagnostics of closure behavior. We will refine this interpretation and strengthen the physical explanation in the revision.
We also acknowledge the importance of clarifying the role and limitations of the ΔS ≈ 0 assumption at annual scales, especially in the context of disturbance-driven shifts in infiltration, recharge, and groundwater contributions. This will be addressed explicitly in the revised manuscript.
Similarly, we agree that further clarification of burn severity treatment and vegetation controls would strengthen the mechanistic interpretation of ET suppression and recovery trajectories. We will expand the methodological description and consider additional integration of vegetation metrics where appropriate.
We appreciate the detailed line-by-line comments and references provided. A structured, point-by-point response will be provided in the formal revision following the discussion phase.
Citation: https://doi.org/10.5194/egusphere-2025-6193-CC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 227 | 127 | 25 | 379 | 36 | 15 | 18 |
- HTML: 227
- PDF: 127
- XML: 25
- Total: 379
- Supplement: 36
- BibTeX: 15
- EndNote: 18
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript presents a comprehensive and well-executed analysis of wildfire impacts on evapotranspiration, basin outflow, and water-balance behavior across major California watersheds. The long-term perspective, integration of multiple datasets, and focus on water-supply relevance make this a valuable contribution to the postfire hydrology literature. The results are generally convincing and clearly presented. I have a few comments below that I believe would help strengthen the methodological clarity and interpretation:
The watershed descriptions highlight volcanic terrain, groundwater-fed baseflow, and snowmelt-driven recharge, suggesting substantial subsurface storage capacity in several basins. While FNF integrates both surface and subsurface discharge, the manuscript does not explicitly discuss potential changes in basin water storage (ΔS). Although I am not deeply familiar with California's montane aquifer and soil storage dynamics, postfire changes in infiltration and recharge could plausibly lead to transient storage effects that influence P–ET–FNF residuals. Clarifying whether storage changes are assumed negligible, and over what timescales, would strengthen the interpretation of the water-balance analysis.
Several analytical thresholds appear somewhat arbitrary and would benefit from additional justification or brief sensitivity testing. These include the 75% recovery benchmark, the 3% burned-area threshold used to define high-fire years, and the 500 m buffer for selecting unburned reference areas. Providing a short explanation or supporting references for these choices would improve methodological transparency.
The CECS ET product is derived from internal water-balance calculations and is then used in the basin-scale closure analysis. Because ET, P, and runoff are therefore not fully independent within this framework, it would be helpful for the authors to clarify how assumptions within the CECS product may influence the observed P–ET–FNF residuals. A brief discussion of potential error propagation or circularity would strengthen confidence in the closure diagnostics.
The manuscript primarily attributes closure deviations to precipitation underestimation and stream-gauge bias, which is plausible given the use of gridded 30 m precipitation data. However, this interpretation remains largely inferential. Comparison with one or more independent precipitation products (where available) could help further support this conclusion and strengthen attribution of closure imbalances.
Most fires affected relatively small portions of the basins, which likely limits detectability of hydrologic responses at downstream gauging stations. While this issue is acknowledged, additional discussion of potential scale mismatch between burned areas, ET aggregation, and FNF stations would help clarify the limits of inference, particularly for weaker runoff responses.