Detailed Response to Reviewer

We would like to sincerely thank the reviewer for the thoughtful and constructive feedback. We greatly appreciate the time and effort invested in providing such detailed comments, which will be very helpful to us improving the clarity, precision, and overall quality of our manuscript.

Reviewer comments are repeated in italics, answers are given with normal letters.

The manuscript "Exploring Controls on Solute Export Mechanisms for Major Nutrients in Anthropogenically Impacted Catchments in Southern Germany under Climate Change" by Frietsch and Schütz (2025) presents results from an analysis of export patterns across seven distinct solutes, measured at the outlets of 40 catchments in southern Germany. This analysis reveals changes in the concentration-discharge relationship of these solutes across two consecutive multi-annual periods, as well as between seasons and varying catchment wetness conditions. These changes imply alterations in the underlying hydrological and biogeochemical processes, which could be driven by climate change, among other influences. The study is grounded in a thorough literature review and extensive analysis that yields interesting and relevant results. However, I am not entirely convinced by the attribution of these results to climate change. Additionally, I found the information about the analyzed catchments to be lacking in detail, and I question the rationale for grouping the solutes into different categories. Finally, while the manuscript is generally well written, it would benefit from an additional grammar check and should be shortened, particularly in the discussion section. Please see more details on my major comments and minor line-by-line comments below.

We agree that attributing observed changes in solute export mechanisms solely to an undefined climate change would not make the point, which we try to argue with our contribution. Locally, climate change is apparent in statistical changes of the occurrence of specific weather patterns and hydroclimatic variables. Namely, it includes by definition the increasing frequency of droughts, more intensive rain events, and increased temperatures (observed very clearly after 2018) as mentioned by the reviewer. In our manuscript, so far we describe climate change as long-term alterations in climatic parameters such as mean annual precipitation, potential evapotranspiration (PET), and actual evapotranspiration (ET) over multi-decadal time scales. These parameters have measurably changed across the study region and within individual catchments (details follow in our answer to major comment 1). Furthermore, climate related controls on solute export mechanisms have been found being significantly correlated with most of the identified changes in solute export patterns. Therefore, the ongoing and widespread shifts in hydroclimatic conditions might have substantial potential to alter environmental systems, including the transport behavior of dissolved nutrients (by clearly acknowledging that changes in the frequency of occurring extreme weather events will influence the statistics as well). Hence, these events act as apparent “agents of climate change” and are amongst others potential means by which climate change is altering the solute export mechanisms within specific catchments.

To reflect the concerns of the reviewer, we will redefine the term climate change and its implications on local weather patterns. Within the discussion we will specifically emphasize the effect of extreme events such as droughts and heavy rainfall events on solute export mechanisms and how they are interlinked with the temporal alteration of mean hydroclimatic conditions and runoff generation processes.

While statistical climate change appears to be a central driver of the observed shifts, we recognize that other factors such as land use, soil properties, legacy effects, and other anthropogenic influences also contribute to system behavior, although they do not represent the dominant or shifting controls in our study. We will revise the manuscript to better reflect this multi-causal interpretation, presenting climate change (and associated weather patterns) as a dominant factor implying alterations in solute export mechanisms, but not as an exclusive factor.

In response to the reviewer’s additional concerns, we will include more detailed information about the analyzed catchments to improve transparency and context. Furthermore, we will rephrase and clarify the terminology used for grouping solutes to improve the rationale behind the categorization.

Major comments

1 Attribution to climate change

As indicated in the title, the goal of the manuscript is to identify the impact of climate change on solute export. However, I am not convinced that this objective is achieved. The authors compare two time periods (before and after 2012), with an unspecified duration for the respective time series, in terms of concentration-discharge (CQ) relationships.

We will specify the duration of the time periods 1+2 in the revised manuscript. They endure from 1979-2003 years before 2012 and to 2012-2022 respectively.

Differences in the CQ relationship between these two periods are attributed to climate change. While I agree that climate change might be a crucial driver for solute export patterns, I find the attribution of the observed effects to climate change unconvincing. In particular, I noted a lack of information on how periods 1 and 2 differ specifically in terms of climatic conditions, as well as how these differences vary among catchments. It remains unclear whether the observed effects are a result of the drought that began in 2018, more intensive rain events, increased temperatures, etc. Although comparing two periods is acceptable (albeit climate change is inherently a gradual process), I strongly recommend conducting an analysis of hydro-climatic anomalies between these periods, accounting for spatial variation, and then attributing these anomalies to the observed changes in CQ relationships rather than to climate change per se. Furthermore, care should be taken in attributing every temporal change to climate change since, for example, Erhardt et al. (2019) demonstrate that CQ relationships can shift over time due to past changes in solute input. Additionally, it should be noted that the period before 2012 is not “unaffected” by climate change, as stated in the caption of Figure 1.

We agree with several of the reviewer’s concerns and will revise the manuscript to substantiate our interpretations more thoroughly. Hence, we recognize the need for a clearer and more detailed justification of this interpretation.

To support our approach, we will present more detailed data on key indicators of climate change, such as increasing air temperatures and the growing frequency of extreme precipitation events in southwest Germany. These trends are documented in the KLIWA climate report (KLIWA, 2021), which we will reference and incorporate into our revised analysis, and further support with absolute values to substantiate our statements. According to the KLIWA report, average air temperature is increasing at a rate of 1.4 to 1.8°C per 90 years. Additionally, maximum one-day precipitation amounts have increased in almost all of Germany during the hydrological winter half-year (up to +33%), and also during the summer half-year (up to +28%), although there are regional decreases as well. Due to this variability, the maximum one-day precipitation trends are not statistically significant; however, an overall increase in peak discharge events is evident for southwest Germany.

Although we acknowledge that climate change is a gradual process, we argue that a noticeable shift in hydroclimatic conditions can be observed between the two study periods. Period 1, while not entirely unaffected by climate change, is considered to be less influenced and therefore serves as a reference for an incremental change. Period 2, by contrast, includes several especially dry years, notably 2018 and the subsequent years, which exhibit stronger and more frequent indications of climate change. We therefore consider it reasonable to distinguish between a less climate affected and a more climate affected period. Accordingly, we will reword the manuscript and soften the causal language.

Fig.1 (see supplement): Changes in air temperature (left) and mean annual sums of evapotranspiration (ET), potential evapotranspiration (PET), and precipitation (P, mm; right) between period 1 (P1) and period 2 (P2). Tmax = maximum air temperature (°C), Tmin = minimum air temperature (°C), Tmean = mean annual air temperature (°C). Boxplots represent climatic data from 40 climate stations located in the respective catchments under study. Asterisks indicate a significant difference of the means within each group (t-test, p < 0.01).

Across the investigated catchments, mean annual air temperature increased significantly from 8.68°C in Period 1 to 9.70°C in Period 2. Potential evapotranspiration (PET) also increased markedly from 578.47 mm to 692.94 mm, and actual evapotranspiration (ET) rose from 460.33 mm to 490.19 mm. While mean annual precipitation showed a decreasing trend from 904.61 mm to 818.72 mm, this decline was not statistically significant. However, the inclusion of extreme years such as 2018 and the following years within Period 2 further underscores the growing influence of climate change.

In the revised manuscript, we will provide both region-wide climatic data and absolute hydroclimatic values for each period to document differences. This information will be presented for the entire study area, as well as for the individual catchments investigated. These additions will be included in the materials and methods section to strengthen the basis for interpreting observed shifts in concentration-discharge relationships in the context of changing climate conditions. Nevertheless, we will additionally account for possible other attributions of the observed changes in in solute export mechanisms in the revised manuscript.

2 Catchment characteristics

From my perspective, the manuscript lacks a table summarizing the catchment characteristics, particularly concerning the variables listed in Table 1. Including such a table would help readers understand the variability across catchments and how these variables may relate to changes in solute export patterns.

We will include a detailed table in the supplementary materials that clearly presents the characteristics of the catchments and the underlying data used to infer the variables appied in this study, helping readers to better understand how the variables relate to changes in solute export patterns.

3 Characterization and grouping of solutes

I am not entirely convinced by the classification of solutes into different groups. The terminology used raises some concerns: while it is evident that nitrogen (N) and phosphorus (P) are major nutrients, also carbon (C) can be conceived as a nutrient (see Wachholz et al. 2023), as well as to magnesium (Mg) and calcium (Ca), which are secondary nutrients rather than primary ones. When comparatively describing these solutes, greater care should be taken not to mix up their functional roles (e.g., nutrition) with their origins (e.g., geogenic sources).

Additionally, the grouping of shallow-sourced & biogeochemically affected, shallow-sourced & discharge-driven, and discharge-driven groundwater-sourced solutes appears unconvincing. Firstly, it is inconsistent with what is illustrated in Figure 5, where nitrate is represented across all layers. Secondly, I would recommend avoiding the term "discharge-driven," as discharge results from various factors, including catchment wetness and hydrological connectivity. Thus, solute export patterns may be influenced by hydrological processes but not necessarily by discharge as such. Furthermore, the definition of "biogeochemically affected" lacks precision. For instance, while nitrate may not be as volatile as ammonium (NH4), it is still subject to various biogeochemical processes, such as uptake and denitrification. Referring to Figure 5 again, total phosphorus (TP) is only depicted in the surface layer, while soluble reactive phosphorus (SRP) is depicted in the subsurface layer as well. It is unclear how SRP could be present in the absence of TP. Overall, I find the reasoning behind the grouping of solutes to be unclear, as well as the justification for its generalizability across the observed catchments.

The reasoning of the reviewer can be followed in several points, and the respective revisions will be incorporated into the manuscript. Whilst the grouping of solutes will remain in its original form as a meaningful basis for analysis, it is acknowledged that the current interpretative labels, such as "shallow-sourced biogeochemically affected solutes", "shallow-sourced discharge-driven solutes", or "discharge-driven groundwater-sourced solutes", may be ambiguous and potentially misleading.. The grouping of the solutes has been based on descriptive data analysis with regard to temporal changes in the respective c-Q relationship, which will be explained in the revised manuscript in detail.

The group of “temporal-dynamic solutes” shows seasonal variability in their c–Q relationships (slope b), while “short-term stable”and” long-term stable” solute c-Q relationships do not change between seasons. . Both,”temporal-dynamic” and “short-term stable” solutes exhibit significant changes in slope b between Periods 1 and 2, indicating shifts in solute export mechanisms. In contrast, “long-term stable” solutes display no significant changes in slope b either seasonally or over time, suggesting temporally consistent export patterns. Hence, the three groups will be renamed as follows:

(a) temporal-dynamic solutes,

(b) short-term stable solutes, and

(c) long-term stable solutes.

In line with the reviewer’s suggestion, the term “discharge-driven” will no longer be used. This reframing emphasizes temporal characteristics rather than mechanistic interpretations, offering a clearer and more objective basis for discussion. The group labels will be revised to reflect observed temporal behaviour, specifically regarding intraannual and interannual variability.

Furthermore, following the reviewers suggestions total organic carbon (TOC) will be incorporated within the nutrient category, rather than being listed separately. To maintain a clear distinction from primary nutrients, calcium (Ca²⁺) and magnesium (Mg²⁺) will continue to be referred to as "geogenic minerals". It should be noted that, when comparing the water quality parameters, the functional roles (e.g., nutritional relevance) and origins (e.g., geogenic sources) will no longer be used simultaneously or in combination.

4 Grammar and manuscript length

The writing quality is generally good. However, the manuscript requires a careful check for grammar, word choice, and consistency, and it should be shortened, particularly in the discussion section. While I appreciate that the manuscript is grounded in a thorough literature review, the discussion would benefit from being streamlined to emphasize the main message. In addition to these three major points, you will find my minor comments below. Please note that I may not have grasped all the details of the discussion, as it is quite lengthy, making it sometimes difficult to extract the main message.

Since it is our intention to convey the main message as clearly as possible to our readers, we will make a concerted effort to shorten the manuscript, particularly the discussion section, to improve clarity and focus.

Minor comments

Title

The term “major nutrients” is misleading, as explained in major comment 3.

We follow the reviewer concerns and change the wording of the title. The new title will be called: “Exploring Controls on Solute Export Mechanisms in Anthropogenically Impacted Catchments in Southern Germany under Climate Change”

Abstract

L9: When I first read "SEM," I initially thought the reviewers were referring to structural equation models, which is a common interpretation of this abbreviation. To avoid confusion, I suggest not using the abbreviation and instead writing it out in full throughout the manuscript.

Some readers might find the term 'SEM' misleading. However, after consideration, we have decided to keep the abbreviation because including the full term would result in too much wording. The frequency of of appearance of the term SEM is high throughout the manuscript, which would make it inconvenient to read it in full every time.

L14: One example of why a grammar check is necessary: “solutes exhibit slightly increase in diluting export mechanisms” should be revised to “solutes exhibit a slight increase in dilution patterns.” (Please note that "dilution" is a pattern resulting from a mechanism, rather than a mechanism itself.) There are more instances requiring correction, which I will not all mention here. Please refer to my general advice above for a thorough grammar check.

As the reviewer remarked, we will refer to 'dilution' solely as a pattern and no longer describe it as a mechanism in itself. Additionally, we will follow the reviewer's general advice and conduct a thorough grammar check of the entire manuscript.

Introduction

L55-56: Here and in other sections, these citations serve as examples of studies that have found similar results. I believe their number could be reduced, and it should be clarified that there are even more examples by adding "e.g."

We will certainly take the reviewer's suggestion into consideration and review the cited literature carefully. Where appropriate, we will reduce the number of citations and eliminate any unnecessary duplication.

L60: Here and in other sections, please ensure that CVC and CVQ are consistently written with subscripts “C” and “Q.”

To maintain clarity and consistency, CV_C and CV_Q will be consistently formatted with subscripts 'C' and 'Q' throughout the manuscript.

Material and Methods

L91-93: There is some redundancy in mentioning the state agencies. This could be one option to shorten the manuscript.

We plan to remove overlapping information about state agencies to make the Materials and Methods section more concise.

L155: I appreciate that b=0 is not automatically attributed to chemostatic patterns; however, I struggle to understand the second part of the sentence, as well as the following paragraph. This may be a wording issue.

We acknowledge the readers' concerns and will improve the wording of the paragraph as follows:

“Further, a slope close to zero (b ≈ 0) suggests that solute concentrations are largely independent of discharge magnitude. However, this does not necessarily imply low variability in solute concentrations (Musolff et al., 2015). In fact, high concentration variability can still occur despite the absence of correlation with discharge. To avoid misinterpreting such near-zero b slopes as indication of chemostatic behavior, we additionally used the CV_C/CV_Q ratio as proposed by Thompson et al. (2011).”

Results

L203-204: This sentence does not convey a clear message. I suggest either expanding it or deleting it entirely and starting with section 3.1.

As recommended by the reviewer, the entire paragraph will be deleted and the revised manuscript will start with section 3.1 directly.

L205-212: I did not observe any trend analysis, and I find it difficult to refer to the differences between the two periods as a “trend.”

The trend analysis has been conducted using linear regression over the entire observation period, however, it has not been documented properly in the method section. It will be documented more clearly as a separate section in the material and methods of the revised manuscript.

L269-270: Or could it be something entirely different?

We acknowledge the reviewer’s comment and recognise that legacy effects or changes in land use and/or land management could be an additional factor influencing changes in SEM. Although, to our knowledge, there have been no major changes in fertilizer regulation in recent years, local legacy effects may occur in individual catchments. We will include the additional information in the manuscript...

L283: What are "surface factors"?

We follow the reviewer and redefine the term "surface factors" by naming the factors directly as follows: “Shallow-sourced solutes, such as major nutrients and TOC, are primarily influenced by near-surface environmental conditions, particularly climate (e.g. temperature and precipitation), as well as soil moisture and the decomposition of organic material on the ground and in the upper soil layers. In contrast, geogenic solutes such as Ca²⁺ and Mg²⁺ are predominantly controlled by subsurface geological factors.”

Discussion

L330: What are fertile sources?

We follow the reviewer's suggestion and will redefine the term 'fertile sources' as follows:

“TOC is less reactive and more persistent, with carbon-rich landscape types serving as sources (e.g. wetlands or riparian zones), visible in higher mean concentrations and broader variability.”

L333: High compared to what? I am unclear about what is meant by “consistent with” in this context. Does this indicate that the concentrations are in a similar range?

We acknowledge the reviewer's concerns about the wording. Thus, the sentence will be revised as follows: “Geogenic solutes, such as Ca²⁺ and Mg²⁺, exhibited mean concentrations comparable to those reported by Musolff et al. (2015).”

L354-356: I disagree with the assertion that shallow sources can lead to both enrichment and dilution patterns. In my understanding, these concepts contradict one another. If a solute has higher concentrations from shallow layers that become connected during high flow, this would cause an enrichment pattern. Conversely, if the source becomes depleted (indicating a limited supply in the shallow layers), this would result in a clockwise hysteresis, with concentrations decreasing even as discharge (Q) is still rising. A dilution pattern would imply that concentrations are higher during base flow (i.e., derived from groundwater or point sources).

We concur with the reviewer's interpretation and acknowledge that our initial explanation of the export processes may have been too unclear. In order to enhance clarity, the paragraph will be revised providing a more precise and detailed description of the implicated mechanisms:

“However, in intensively managed catchments, homogeneously and widely distributed solute sources (e.g., SRP and TP) mask biogeochemical effects, resulting in chemostatic behavior where solute mobilization is proportional to discharge (Ali et al., 2017; Basu et al., 2011; Thompson et al., 2011). Shallow-sourced solutes exhibit chemodynamic enrichment behavior when heterogeneously distributed sources in the upper soil layers are mobilized unevenly as discharge increases. In contrast, dilution behavior can occur when sources in deeper soil layers, which dominate during low-flow conditions, contribute proportionally less during high-flow events, while low-concentration surficial sources become more dominant (Basu et al. 2011; Ebeling et al., 2021; Rose et al., 2018).”

L358: Why only mimicking? There might also be real point sources.

We used the term 'mimicking' deliberately to distinguish between temporally variable biogeochemical processes, such as the biological release of SRP during periods of low flow in summer, and persistent point sources, such as effluents from wastewater treatment plants (WWTPs). To clarify this distinction, we will rephrase the sentence as follows: “Dilution dynamics can also result from the release of biological SRP during periods of low flow, acting as temporal point sources in sediments and riparian zones (Dupas et al., 2018; Ebeling et al., 2021; Smolders et al., 2017).”

L390: I assume you are referring to Winter et al. (2022) here? This fits well and should definitely be cited here, but the year is missing, and the reference is not included in the reference section.

We agree with the reviewers' suggestion and will correct the citation to properly refer to Winter et al. (2022), including the full reference in the reference section.

L506: How can wet conditions be explained by a high drought index?

We understand that this might be confusing. As explained in the Materials and Methods section, higher dMI values indicate more humidity, while lower values indicate dry conditions. To avoid confusion, we will change the term "high drought index" to "humidity inferring drought index values" in this text.

L507: How do arable land and evapotranspiration create internal sources and transport limitations?

We acknowledge that the correlation between arable land, evapotranspiration, and their role in generating internal sources and transport limitations may not have been sufficiently clear. To address this issue, the paragraph will be rephrased providing a more comprehensive explanation of the relevant mechanisms, thus assisting readers in comprehending the argumentation more easily. The following corrections will be applied to the paragraph:

“In contrast, arable land and high evapotranspiration promote enrichment behaviour, respectively. Fertilizer use on arable land leads to nitrogen accumulation in the soil, creating diffuse sources, while high evapotranspiration causes dryness in upper soil layers, reducing constant drainage and by that prevents source depletion. When hydrological connectivity is temporarily restored (e.g., during rainfall events), these accumulated/stored nutrients can be rapidly mobilized, resulting in pronounced concentration increases. Urban areas, by contrast, often show dilution behavior due to relatively constant contributions from point sources (Aubert et al., 2013; Basu et al., 2010; Dupas et al., 2018; Ebeling et al., 2021; Musolff et al., 2015).”

L545: I do not see this as a consequence. A solute concentration driven by hydrological and biogeochemical processes does not mean that its (anthropogenic) input is no longer relevant. These processes are not mutually exclusive.

We are in agreement with the reviewers' evaluation that hydrological and biogeochemical processes influencing solute concentrations do not exclude the relevance of anthropogenic inputs. To reflect this more accurately in the manuscript, a revision of the paragraph is necessary. The following updated text will elucidate that export mechanisms are shaped by natural processes, while human influence remains important for all water quality parameters: “Under prevailing fertilizer application and land manage strategies, the export mechanisms of major nutrients are shaped by a combination of biogeochemical processes and hydrological connectivity, except for NO₃-N. NO₃-N is buffered by large and persistent sources, which reduce seasonal variation and limit changes in solute export due to global warming, whereas nutrients such as NH₄-N, SRP, and TP exhibit weaker buffering effects. However, the influence of mankind remains present for all water quality parameters.”

Figures and Tables

Figure 1: This appears to be a strangely skewed projection of Germany. This should be checked, and the projection type should be indicated on the map or in the figure caption.

We agree with the reviewer´s observation and will apply a more commonly used map projection. The projection type will also be indicated in the figure caption for clarity.

Figure 2: My concerns regarding the grouping definitions have been noted above. Furthermore, I recommend adding labels (a-c) to the panels and ensuring consistent axis limits across all plots to make the differences clearer. Additionally, incorporating areas (e.g., boxes) to indicate chemodynamic and chemostatic patterns, as well as enrichment and dilution, similar to Musolff et al. (2015), might help convey the message more effectively.

With regard to the grouping definitions, we direct the reader to our response to Major Comment 3, where we provide a detailed explanation of our rationale for maintaining the current grouping approach. With regard to the visualisation of export behaviour (e.g., enrichment, dilution, chemostatic, and chemodynamic patterns), these are indicated by the vertical and horizontal reference lines in the figure. This subject will be addressed in the figure caption to enhance the clarity and intuitiveness of the illustration. In addition, panel labels (a–c) will be incorporated.

To avoid what I consider arbitrary grouping, I suggest either plotting all solutes in one graph, displaying each solute individually (with consistent axis limits), or grouping them into categories that require less interpretation, such as N-based (NO3 and NH4), P-based (TP and SRP), TOC, and geogenic sources (Ca and Mg), for example.

We acknowledge the reviewer's point. Nevertheless, we consider the current clustering to be meaningful (see major comment 3).

Figure 3: The y-axis label is not clearly attributed, and the text is relatively small. Consider rotating the figure for better visibility.

In order to enhance the clarity of the figure, the font size will be increased. Furthermore, the meaning of the y-axis label in the figure caption will be addressed as follows: “b = solute export mechanisms (b < 0: dilution; b > 0: enrichment behaviour). Regarding the suggestion to rotate the figure, this modification will not be implemented, as the quality of the figure in the final manuscript will be significantly higher than the version used during the review process. We therefore consider rotation unnecessary.

Figure 5: I suggest testing if these changes are significant. Because if not, the changes should be depicted as equally sized arrows (i.e., indicating no change in pattern) in Figure 5, which may be applicable for Ca and Mg.

We are aligned with the reviewer's perspective. In Figure 5, only statistically significant changes will be represented, as opposed to displaying tendencies. The significance of the data had previously been tested using the Kruskal-Wallis test, and the results of this test are also documented in the supplementary figures. This adjustment is intended to ensure that the figure exclusively highlights robust patterns of change, thereby addressing the reviewer's concerns regarding elements such as Ca and Mg.

Figure S1: Mg is missing

Due to the very similar behavior of Mg²⁺ compared to Ca²⁺ and initial space constraints, Mg²⁺ was originally not included in the figure. However, since space limitations are less critical in the supplementary material, Mg²⁺ will now be added as an additional panel in Figure S1.

Table 4: This table conveys a wealth of interesting information. However, I do not see the triangles mentioned in the caption. Additionally, I suggest adding asterisks to indicate where slopes are significantly different and including the respective number of catchments for each change class (a-c). Alternatively, categories a-c might be combined into a more general message (optional suggestion!). For example, ammonium (NH4) shows a variety of slopes for the black line but comparable red lines, all indicating a higher slope. For soluble reactive phosphorus (SRP), categories a) and b) appear very similar and could be presented as one image. Total phosphorus (TP) consistently shows a higher red slope, while TOC and category c are also quite similar. Nitrogen (NO3-N) and geogenic minerals show virtually no change in red and black slopes.

The impacts of controls (triangles) will be added behind the control labels in the table to improve clarity. We will also review the significance of the slopes and indicate them accordingly. Regarding the grouping, we would like to keep the division into overall catchments, pre-enrichment catchments, and pre-dilution catchments, as this structure effectively highlights differences in export behavior among catchments in Period 1 and their subsequent changes. Additionally, we will include the number of catchments considered in each analysis to provide further transparency.

Acknowledgements

“Further, the data that support the findings of this study are available from the corresponding author upon reasonable request.” – I would prefer to see this information in the supplement. Additionally, I would like a supplementary section that contains detailed information on the catchments, their characteristics (particularly concerning the variables mentioned in Table 1), and the differences between periods 1 and 2.

Upon reconnection with the relevant state agencies, we have confirmed that the data used in this study cannot be freely distributed by the authors. The datasets are available through the responsible authorities and can be accessed upon a justified request. We will clarify this in the manuscript and refer to the appropriate data sources. Additionally, we will include a supplementary section with detailed information on the catchments and their characteristics.

Detailed Response to Reviewer

We would like to thank the reviewer sincerely for their valuable and constructive feedback. These comments will help us improve the statistical rigor, clarity, and overall scientific quality of our manuscript.

Reviewer comments are repeated in italics, answers are given with normal letters.

Summary: The study analyzes concentration-discharge (cQ) relationships for major nutrients (NO₃-N, NH₄-N, SRP, TP), total organic carbon (TOC), and geogenic minerals (Ca²⁺, Mg²⁺) across two time periods: Period 1 (pre-2012) and Period 2 (post-2012). The authors use cQ slopes and coefficient of variation ratios (CV_C/CV_Q) to characterize solute export mechanisms and identify climate change impacts. Key findings include enhanced enrichment behavior for shallow-sourced solutes since 2012, slight increases in dilution mechanisms for geogenic solutes, altered solute source distribution and hydrological connectivity due to climate change, and a shift in controls on SEMs toward greater influence of climatic factors. This manuscript addresses an important question with a substantial dataset but suffers from significant methodological limitations that undermine the scientific quality. Revisions must address fundamental methodological concerns to enhance its scientific contribution.

Strengths: This manuscript addresses a critically important and timely research question by examining how climate change affects nutrient export mechanisms in watersheds, with significant implications for water quality management. The study's most significant strength lies in its comprehensive dataset encompassing 40 catchments across southern Germany with 8-20 years of data, providing robust regional representation and statistical power to detect meaningful trends. The multi-solute approach examining both reactive nutrients and geogenic minerals offers valuable mechanistic insights, while the temporal comparison framework using 2012 as a breakpoint represents a novel analytical approach with clear practical relevance for water resource management.

We thank the reviewer for the thoughtful summary and acknowledge the important methodological concerns raised. We agree with several points and will revise the manuscript to address these issues and strengthen the overall scientific quality. In particular, we will provide a clearer justification for the chosen temporal division (January 1, 2012), supported by additional climate data and statistical analyses to demonstrate significant environmental differences between the periods. To improve the statistical robustness of our results, we will include measures of uncertainty to better assess the influence of climate change on catchment controls. Overall, we are confident that these revisions will address key concerns and enhance the clarity, rigor, and scientific contribution of the manuscript.

Areas for improvement:

Methodological Issues:

The choice of January 1, 2012 as the temporal division point lacks strong scientific justification, as the authors cite KLIWA (2021) climate data but fail to demonstrate that this specific date represents a meaningful threshold for biogeochemical changes. A more robust approach would employ change-point analysis or demonstrate statistical significance of the 2012 breakpoint rather than relying on an arbitrary division. Additionally, the manuscript suffers from inadequate statistical rigor, lacking confidence intervals for cQ slopes, statistical tests for differences between periods, assessment of temporal autocorrelation, and power analysis for detecting changes.

We agree with several of the reviewer’s concerns. We will revise the manuscript regarding the choice of January 1, 2012 as the temporal division point, as also mentioned in our reply to RC1. To provide a more detailed justification for this choice, we will present additional data on key indicators of climate change, such as the increase in air temperatures (see reply to RC1). Moreover, differences in various environmental factors, such as temperature, which support the comparative periods with January 1, 2012 as the division point, will now be substantiated with statistical significance (see reply to RC1). Further, to ensure the robustness of our analyses, we required a sufficiently long and continuous record in the second period. Selecting 1 January 2012 as the start date enables an observation period of 10 years, maximising the stability and reliability of our results. The choice of temporal division point will be explained more clearly in the methods section and its implications will be critically reflected upon in the discussion.

To strengthen the statistical rigor of the analysis, standard errors for all calculated cQ slopes have been added, and error bars are now included in Figure 2 (see revised figure in supplements to the comments). This enhances a clearer representation of the variability in concentration–discharge (c–Q) relationships within individual catchments.

Differences in cQ slopes between the two periods are now formally assessed using analysis of covariance (ANCOVA), allowing for the statistical comparison of regression slopes between Period 1 (before 2012) and Period 2 (since 2012). In the revised visualizations of figure 4, significant differences between periods are highlighted through increased symbol size to aid interpretation.

The influence of catchment descriptors (explanatory controls) on cQ slopes has been evaluated using Pearson correlation tests. To account for multiple comparisons across 23 variables, p-values have been adjusted using the Benjamini–Hochberg correction. In addition, analysis of covariance (ANCOVA) has been conducted to test for significant differences in the relationships between explanatory controls and cQ slopes across the two time periods, and to examine whether the influence of explanatory controls has shifted over time. Updated results are presented in Table 3 (see supplements to the comments). Non-significant correlations highlighted in red will be removed. Further, the table description will be revised to include detailed statistical information, such as significance thresholds and adjustment methods, in order to enhance clarity and transparency in the interpretation of results.

These methodological refinements provide a more rigorous foundation for interpreting differences between time periods and for evaluating the role of potential drivers of change. All statistical corrections, including the use of adjusted p-values (Benjamini–Hochberg procedure), the application of ANCOVA for slope comparisons, and the inclusion of standard errors for cQ-slopes, will be transparently addressed in the newly revised Methods section.

Data Quality and Sampling Issues:

The authors acknowledge that their monthly to biweekly sampling frequency may miss critical high-flow events where nutrients are preferentially transported, potentially underestimating loads and fundamentally undermining their conclusions about changing export mechanisms (Lines 325-329). While 40 catchments provide reasonable coverage, the selection criteria (Lines 119-132) may introduce bias toward certain catchment types, limiting the broader representativeness of the findings.

We acknowledge the concern regarding the lower sampling frequency (monthly to biweekly), which may potentially miss critical high-flow events that can significantly contribute to nutrient export. However, large-scale monitoring networks seldom maintain higher frequencies, so our analysis necessarily relies on existing monthly to biweekly records. Thus, we optimized our analysis to extract the most reliable insights from the available datasets.

This sampling frequency allows us to capture seasonal variability in solute transport, with higher flows in winter and lower flows in summer. However, we acknowledge that event-based transport processes, such as those described by Knapp et al. 2020 and Winter et al. 2022, cannot be resolved at this resolution. These short-term events may amplify or counteract the seasonal transport patterns, especially for more dynamic solutes such as NH₄⁺, SRP, TP, and TOC. In contrast, more conservative parameters like Ca²⁺ and Mg²⁺ are less sensitive to such event-driven dynamics due to their greater long-term stability.

We are convinced that hydroclimatic events, especially prolonged droughts, are associated with climate change and play an important role in our study. This ensures more frequent sampling during low-flow conditions. Crucially, the sampling interval remained unchanged between the two periods, preserving comparability. Therefore, these potential effects do not undermine our conclusion that the overall trends are attributable to climate change, as trends were detected despite constant sampling procedures (sampling interval). Nevertheless, we will consider the potential effects of missing high-flow events in the revised manuscript. The discussion will emphasise that the study focuses on seasonal solute transport patterns while acknowledging that event-based processes could not be analysed due to constraints in the data resolution. We will also highlight the relevance of event-based processes as a promising avenue for further investigation, to quantify how baseflow‑focused sampling under drought conditions may influence long‑term water‑quality assessments.

Further, the selection criteria may introduce bias towards certain catchment types. However, these criteria are essential to ensure that concentration (c) and discharge (Q) measurements are directly coupled, which presents a significant challenge in Germany, where c and Q are usually monitored in different locations. Enforcing these coupling rules optimises the use of the public database while preserving the integrity of the analysis. Nevertheless, the selected catchments still represent a diverse range of characteristics, including variations in size, altitude and geology, ensuring that the findings remain broadly informative.

Climate Attribution Problems:

The study fails to adequately separate climate change effects from other confounding factors, including changing agricultural practices, policy interventions such as EU Water Framework Directive implementation around 2000-2015, and other environmental changes that could influence nutrient export patterns. The manuscript relies heavily on KLIWA (2021) for climate characterization but doesn't provide detailed climate trend analysis for their specific study period and locations, weakening the attribution of observed changes to climate drivers.

Although agricultural practices, policy measures (e.g., the EU Water Framework Directive), and other environmental changes can influence nutrient export, our results indicate that land-use effects on SEM remained stable across both periods. While the influence of EU policies cannot be entirely excluded, the SEM analysis revealed significant correlations primarily with climatic, soil, and geological variables, whereas landscape factors like arable land showed limited influence and no change between the observed periods (see table 3). Accordingly, agricultural factors are considered to play a subordinate role in this context. Nonetheless, their potential effects will be addressed more thoroughly in the revised discussion section.

Furthermore, for each observation period, mean climatic descriptors were calculated individually for each catchment. Specifically, we used the periods 1982/1991–2011 and 2012–2022 to derive mean climatic variables, allowing us to account for long-term trends and differences in local climate conditions. This approach strengthens the attribution of observed changes to climate drivers by explicitly considering catchment-specific climate developments alongside land-use and landscape controls (see comments to RC1).

Conceptual and Analytical Issues:

While the authors propose mechanisms for observed changes (Lines 430-443), many explanations remain speculative without supporting process-level data such as soil moisture or plant uptake measurements to validate their hypotheses. The study examines catchment-scale responses but infers field-scale biogeochemical processes without adequate mechanistic support, creating a problematic scale mismatch between observations and interpretations.

This study uses extensive databases, which include data provided by governmental authorities as well as publicly available datasets. The focus of our analysis is to evaluate these existing data, which also include variables such as soil moisture and to identify differences and relationships through advanced data analysis approaches. We rely on comprehensive, spatially resolved datasets from the German Weather Service (DWD), which provide high‑quality, area‑wide coverage. Soil moisture, for instance, was calculated for each catchment as the percentage of plant‑available water (% nFK), assuming a sandy‑loam texture with a field capacity of 37 vol%. These reliable datasets ensure robust and consistent input for all analyses.

Further, the study is based entirely on extensive database analysis and does not include new field measurements, which would be impractical on a regional scale due to resource constraints. Instead, we apply well-established biogeochemical interpretations from the literature to our datasets. We accept the reviewer’s criticism, and in the revised discussion we will highlight more clearly that our biogeochemical interpretations are grounded in previously published findings. Many of these studies employed comparable methods and similarly structured datasets. Aligning our interpretations with these established studies, we ensure a robust contextual foundation for assessing biogeochemical processes and extending them to evaluate the influence of climate change.

Specific line items (Specific Technical Issues):

Lines 64-67:

Run-on sentence needs restructuring for clarity.

We will rewrite the sentences as follows to improve clarity: “The combined approach of cQ-relationship and CV_C/CV_Q exhibits temporal variability in solute concentrations and can identify flow conditions with elevated solute levels. High solute concentrations are linked to eutrophication processes that harm aquatic ecosystems and pose risks to drinking water quality (Halliday et al., 2013; Radach et al., 2010; van der Velde et al., 2010; Winter et al., 2020).”

Lines 154-157:

The definition of chemostatic vs. chemodynamic behavior using CV_C/CV_Q > 0.5 threshold needs better justification. This threshold appears arbitrary.

We acknowledge the reviewer’s concern regarding the choice of the CV_C/CV_Q > 0.5 threshold to distinguish chemostatic from chemodynamic behavior. Thompson et al. (2011) demonstrated that the CV_C/CV_Q ratio for a conservative tracer was consistently around 0.5, with variability in concentration mainly driven by stochastic inputs in the recharge. We use this value as a benchmark to clearly identify chemostatic conditions representative of conservative cQ behavior. This approach is also supported by Musolff et al. (2021), who applied the same threshold to analyze spatial and temporal variability in concentration-discharge relationships.

We acknowledge that threshold values in the literature vary (e.g. 0.5 or 1). In our study, a value of 0.5 aligns with the behaviour of the solute groups, enabling a clearer distinction between chemostatic and chemodynamic patterns. This threshold is used solely as a structural tool to categorise and distinguish the solute groups. A higher threshold (e.g. 1) would be less appropriate for our data and impair the distinction between the solute groups.

Lines 208-212:

The reported concentration decreases (60-73% of catchments) are substantial but lack statistical significance testing.

The trend was tested using linear regression, and significance was evaluated using the F-test from ANOVA and additionally corrected by Benjamini-Hochberg correction (p<0.05). Results are applied on Table 2 (see supplements on comments).

Lines 223-228:

Repetitive description of solute behavior.

To address this comment and improve the manuscript’s readability, we will remove the repetitive description of solute behaviour to shorten the publication.

Lines 285-297:

The correlation analysis between catchment characteristics and cQ slopes needs correction for multiple testing and should report effect sizes, not just significance.

To ensure a more robust assessment of statistical significance, the Benjamini-Hochberg correction will be applied considering 23 variables. In addition, Pearson's r will be included to provide a measure of effect size (see table 3 in supplement to the comments).

Lines 477-482:

The mechanistic explanation for NH₄-N behavior is overly simplistic and doesn't account for complex nitrogen cycling processes.

The aim of this study is not to explore detailed biogeochemical processes in depth. Instead, it simplifies these processes and focuses on identifying the sources and transport pathways of NH₄-N. Our findings align closely with literature documenting biogeochemical processes, and have been simplified for rough interpretation on a large scale.

Table 2:

Standard deviations should include sample sizes.

Sample sizes will be included to improve statistical transparency for each parameter (see table 2 in supplements to the comments).

Table 3:

Correlation symbols are inconsistent and poorly explained.

Information about the explanatory variables and related variables will be added to the description to improve clarity and understanding (see supplements to the comments). Correlation symbols in the text (l. 285-297) and supplements will be adjusted for greater consistency; no further inconsistencies were identified.

Figure 1c:

The cQ-relationship comparison is visually compelling but lacks statistical analysis of differences between periods.

We appreciate the reviewer’s observation regarding the lack of statistical evaluation accompanying the visual comparison of cQ relationships. To address this, statistical differences between Period 1 and Period 2 have been assessed using analysis of covariance (ANCOVA). For example, in Figure 1c, the difference between periods in the Fils catchment is statistically significant (p < 0.001), as determined by ANCOVA. This information will be incorporated into the figure description to provide greater analytical clarity and support for the visual interpretation.

Figure 3:

The scatter plots showing temporal differences would benefit from regression lines and confidence intervals.

Regression lines were initially omitted because temporal differences for overall, dilution, and enrichment catchments are already evaluated in table 4. However, to improve clarity, regression lines will be added to figure 3 where significant temporal differences occur (see supplements to the comments). To maintain visual clarity, we refrain from adding confidence intervals, but the regression lines clearly deviate from the 1:1 line, underscoring relevant shifts between the two periods.

Figure 5:

The conceptual diagram oversimplifies complex biogeochemical processes and transport pathways.

The aim of this study is not to explore biogeochemical processes in detail. Instead, the conceptual diagram illustrates dominant transport pathways at the catchment scale and provides a general overview of how these pathways may change in response to climate change. As we will explain in the discussion, the diagram is a necessary simplification that highlights general patterns and dominant processes. Due to the resolution of the underlying data, event-based dynamics and fine-scale biogeochemical mechanisms cannot be captured in detail (see Knapp et al., 2020; Winter et al., 2022; Data Quality and Sampling Issues). Nevertheless, interpretation is guided by well-established findings from the existing literature. Accordingly, the focus remains on illustrating broad-scale transport behaviour rather than event-scale processes, providing a general overview of how these pathways may shift in response to climate change.

Throughout the manuscript:

Results lack uncertainty quantification.

We appreciate the reviewer’s concern and agree that adding the above-mentioned uncertainty quantification would improve the clarity and statistical rigour of our results. In this context, we will conduct different statistical tests and present and describe the tests we have already performed more clearly in the revised manuscript.

References:

Musolff, A., Zhan, Q., Dupas, R., Minaudo, C., Fleckenstein, J. H., Rode, M., Dehaspe, J., Rinke, K. (2021). Spatial and Temporal Variability in Concentration-Discharge Relationships at the Event Scale. Water Resources Research, 57(10). doi: https://doi.org/10.1029/2020WR029442

Thompson, S. E., Basu, N. B., Lascurain, J., Aubeneau, A., and Rao, P. S. C.: Relative dominance of hydrologic versus biogeochemical factors on solute export across impact gradients, Water Resources Research, 47, 10. doi: 10.1029/2010wr009605, 2011.

Winter, C., Tarasova, L., Lutz, S.R., Musolff, A., Kumar, R., Fleckenstein, J.H. (2022). Explaining the Variability in High-Frequency Nitrate Export Patterns Using Long-Term Hydrological Event Classification. Water Resources Research, 58. doi: https://doi.org/10.1029/2021WR030938

Knapp, J.L.A., von Freyberg, J., Studer, B., Kiewiet, L., Kirchner, J.W. (2020). Concentration–discharge relationships vary among hydrological  events, reflecting differences in event characteristics. Hydrol. Earth Syst. Sci., 24, 2561–2576. doi: https://doi.org/10.5194/hess-24-2561-2020