Nutrient Load Mitigation with Wintertime Cover as Estimated by the INCA Model

Round 1

Reviewer 1 Report

The article describes the results of modelling the influence of agricultural management on the concentration of nutrients in the runoff from a Finish catchment with mixed land-use (arable land 28%, forest 57%, mires 5%, residential area 5%) using the INCA model. After calibration, the model was used for a scenario study to show how agri-environmental measures can affect phosphorus and nitrogen exports from the catchment. The literature review in the Introduction of the article well reflects the current state of knowledge both in general and in the described catchment and emphasizes the problem of agri-environmental measures based on no-till techniques, which can efficiently reduce erosion and losses of particular P from arable land areas to surface waters, but can increase export of dissolved P, causing eutrophication in the receiving water bodies. The results of this modelling study are in-line with a common view of the effect of no-tillage field management, i.e., they indicated an increase in bioavailable P in the surface water runoff, although total P and total N were reduced. However, despite these rather reasonable conclusions, I have a number of reservations about the description of data and use of the model that should be resolved and improved before the publication of the article, namely:

• In fact, there is no description of the observed data that was intended to simulate, except for one graph, which shows just a short segment (6 years out of a total of 22 years) of observed TP concentrations together with the modelled results. The observed data of TP, TDP, PO4-P (DRP) and PP should be presented, together with an assessment of their time trends and seasonality. It would also be good to show other potential drivers of changes in phosphorus concentrations in the catchment runoff, which may include, in addition to the presented changes in the share of no-tillage fields, changes in livestock numbers, wastewater discharges, hydrology, climatic conditions, etc. For example, it is known that with a decrease in atmospheric acidic deposition, increasing temperature and greater fluctuations in precipitation in recent decades, the concentration of DOM in the runoff from forest catchments in temperate regions, including Scandinavia, increases, and with it also the concentration of organic P. Therefore, the authors should also take this into account when analysing changes over time.
• The calibration of the INCA-P model for TP and PO4-P was completely wrong. According to the results shown in the table and graph (only for TP; Fig. 3), the model was not able to reproduce changes in concentrations during the season, only very approximately (with a relative error of 20 to 25%) hit the averages of the observed data. From the only graph shown for TP (Fig. 3), it can be seen that the model generated many false peaks. Apparently, it did not correctly describe the processes involved in the runoff of P from the catchment, which is in conflict with statements in conclusions. The major outcomes of the study are based on these results, which is very unreliable. In the modelled scenarios, the changes in the average values are in the range of up to 10%, which is less than the relative error of the modelled average during calibration.
• It is not clear from the text why the results of the model were so bad: whether it is the inadequacy of the model itself or some specificity of the modelled catchment that was not well captured in the concept of building the model. There is a lack of information and data in the article, in particular: (i) how the outflows from the forest, which makes up the majority of the catchment area, have been assessed and calibrated; (ii) boundary conditions for the operation of the model are not specified, especially those that affect the export of P, i.e. inputs to the river basin by fertilization with mineral and organic fertilizers, atmospheric deposition, wastewater, etc .; (iii) in the results describing the sensitivity analysis (Fig. 8), the parameters of the model are given only in general terms and it is not possible to determine exactly which one was involved; (iv) information on model parameterization in sub-basins is missing; this should be included at least in the Supplementary information; (v) missing presentation of simulated vs. observed data TP, PO4-P (DRP), PP in the whole simulated series, as well as the analysis of trends of observed and simulated data and their seasonality, which could help to reveal problems in the model and its calibration.

Specific comments:

L 19-23: The text starting with „We concluded ...“ is not based on results of this study. The poor model performance contradicts to these statements.

L 34: Replace „delivered from“ with „as shown at“.

L 36: Remove „again“; a comma can be added after „area“.

L 45-49: The content of this paragraph should be supported by a reference to the literature.

L 77: Replace „surface cover“ with „no-till fields“ or something similar.

L 203-2011: The paragraph on RUSLE can be omitted because the use of RUSLE is not the primary subject of the article. The references included here should be moved to the text on L 259-263.

L 324: Change „-Shutcliffe“ to „-Sutcliffe“

Figs 4, 5, 6: Font should be increased.

Fig. 7: There is an apparent discrepancy in P-isotherms that were observed for catchment soils to that used in the INCA-P model. This should be explained and discussed in the text.

L 431-433: I cannot agree with advocating bad model results by stating low values of the average relative error. The agreement of the averages of the modelled and observed values with the variability of maximum 17% explained by the model has no significance. In addition, it is not true that the "observed values lied between simulated uncertainty bounds" in Fig. 3. At least half of the observed points lie outside.

L 459-460: I strongly disagree with the statement "We concluded that INCA-P was able to describe both fast, immediate processes but also slow, non-immediate processes that influence P loading from catchments." The results presented in this article do not support any such conclusion.

Author Response

Calibration/validation of INCA-P

Our motivation to concentrate on fields and agri-environmental measures is based on our previous paper, where we concluded that changes in general land use, animal density, point sources or climate are not enough to explain changes in water quality (Rankinen, K., Gao, G., Granlund, K., Grönroos, J., Vesikko, L. (2015) Comparison of impacts of human activities and climate change on water quantity and quality in Finnish agricultural catchments. Landscape Ecology 30, 415-428.). We have added more discussion of that on Lines 130- 132 and 518-26

Yläneenjoki catchment is very erosion sensitive area due to high slopes and erosion prone (Ca poor) clay soils that were developed under sea phases after Ice age. Most of the P is adsorbed on soil particles, and thus peaks in suspended sediment concentrations are reflected as peaks in TP concentration. High variability poses challenges to drawing clear and illustrative figures. We added Figures 3 and 4, which show variability and seasonality of TP and DRP concentrations. PP observations we do not have, and TDP follows the pattern of DRP, but is a bit higher.

Influence of forest areas is probably reflected on TDP concentration which is higher than DRP concentration. That may be P in organic matter, but the INCA-P model do not have specific organic P pool (L 254). Calibration of forests are added on Line 304.

Figure of uncertainty analysis was change to be in color, and parameter names was added there. More discussion of resultas is on Lines 511.

Reviewer 2 Report

Nutrient load mitigation with wintertime cover as estimated by INCA-P By Rankinen, et al. General comments: After reviewed this manuscript, I found the research topic quite interesting. The data analysis method is well-selected and the results are properly interpreted. The quality of this work is high. I enjoyed the style of the writing and data interpretation in this manuscript. I have no specific question about the manuscript and would like to suggest the acceptance of this manuscript for publication in Water.

Author Response

Thank you very much

Reviewer 3 Report

This study evaluated 1) the reduction of P and N loads to receiving waters in response to the European Union’s agri-environmental measures using process-based modeling and 2) the performance of INCA-P model for simulating P dynamics in an agricultural catchment. The article is in line with the journal’s aims and scope, and it presents sufficient information about the research. In my opinion, only minor revision is required.

• The authors should check for typos and grammatical errors. Most of the units must be revised. For example, in Line 68, the unit kg ha-1 should use superscript for ^-1; in Line 112, the sentence is not complete with “(e.g.”; in Line 183, r2 should be R². Be sure these mistakes are eliminated in the next submission.
• Line 314: Scenario 4 is missing. Or there is no Scenario 4? Please clarify.
• What statistical analysis was used in this study?
• Figure 4: The lines are not easy to distinguish. Please use different line types or use color.
• The performance of INCA-P model for simulating P dynamics was not sufficiently discussed. Please add an additional section to discuss it.
• The structure of the article can be rearranged to better link the discussion to the results and goals of the study.
• Since it’s a long article, a brief conclusions section would benefit.

Author Response

More description of INCA-P model structure and calibration are added on lines 254 and 304. In addition, we added new figures (Fig. 3 and 4) to show the level and seasonality of observed and simulated concentrations.

We corrected typos and improved the quality of figures. Scenario 4 did not exist, it was mistake.

We collected conclusions to one paragraph.

Reviewer 4 Report

Water, Peer Review 1068224

“Nutrient load mitigation with wintertime cover as estimated by INCA-P”

General comments:

The paper submitted for review is an interesting case study of INCA model as a tool for predicting the dynamics of nutrient runoff from the catchment area with a changing management method. The work uses impressive data resources, which allowed for free work on studying various mathematical models. The authors present the results of these analyzes pointing to the great potential of the INCA model. The results of the work are presented quite clearly, although in my opinion some aspects require clarification. In general, the work is worth publishing, but there are a few corrections that I present below.

Title

Please consider whether it is appropriate to use the term INCA-P if the nutrient load (both P and N) is included.

Abstract

The abstract is clearly written, stating the purpose, general methods, and results of the study.

Introduction

This chapter deals with the problem of the impact of nutrients on the aquatic environment. The most important thread, further developed in the work, is the phenomenon of the increase in DRP charge from no-till fields compared to the cultivated fields.

In the sixth paragraph the authors synthetically indicate the nature of this phenomenon, but in my understanding the thread is so important that it should be explained in more detail. In particular, the sentence "[13] estimated P balance ..." (line 67-71) is unclear (and too long). At this point, we should describe this phenomenon in more detail, in particular explaining its causes. Does the mentioned 3 kg / ha correspond to the actual fertilization needs in no-till fields?

l.40 The closing brace is missing

l.42. It is not a rule that small lakes are nitrogen limited and large lakes are phosphorus limited. It depends more on the hydrological system of the lake than on its size, because the reason is the drainage of the catchment area. Please consider editing this statement.

Material and Methods

This chapter is the strength of the work. Research methods are described very accurately and logically.

l. 202. Whether the name of subsection 2.3 really should include the phrase "model set ups"? - please note the name of subsection 2.4

l. 314 Please, remove point 4.

Results

l. 381. It requires clarification whether the authors analyzed the results of the river sediment research. There is no mention of this in the work so far.

l. 383-387 Figure 5 shows the concentration of DRP in the runoff waters from the fields. But was the amount of water taken into account in variants a), b), c)? The most appropriate way to assess the expenditure of matter from the studied areas is to address the problem of charge, and not just concentration. A similar remark for fig 6. 

l. 395-396 Please specify the name of the Y axis. In the drawing description, the symbols Ap, BW1, 3C need explanations.

l. 404-405 Figure 8. The current Y axis description contains only 4 parameters: catch, flow, const, soil. Please complete or clarify.

Discussion and Conclusions

l. 432 What figure is it? Unfortunately, I can't see it.

l. 459-460 Isn't that a bit exaggerated? It does not follow from the previous paragraph.

l. 465 - 493 This subsection synthesizes the work of the authors presented in the reviewed article. I admit to feeling a bit uneasy after reading it. First of all, I would be far from generalizing the real impact of the change in the type of field use towards AEM as the obvious cause of the increase in the DRP load in river outlets. The authors cite literature sources which also make conclusions on the basis of modeling. As they indicate - there is no perfect model. These are difficult issues and generalizing on such a key topic as AEM implementation requires extreme caution.

For example - It seems that an important aspect, not taken into account in your research, is the release of phosphorus from the fields after switching to no-till management as a result of a gradual decrease in earlier over-fertilization.

l.502-503. Recommendation without evidence of effect. How do you imagine this plowing? How often? Also in fields with a clear slope? For what permanent grasslands will it be environmentally justified? Your statement is rather a contribution to research aimed at such a solution as a potential method to improve the no-till effect - to be checked and analyzed, rather than a recommendation for implementation.

Author Response

We changed the Title to be more general

Some sentences were rewritten and clarifications are added to the text, e.g. lines 70-80 and lines 581-586. In addition, line 42 is edited to be more general.

We have also added new figures 3 and 4, which show level and seasonality of observed and simulated P concentrations in the river water. When calibrating river water and sediment equations, we used equilibrium isothermes measured by Jukka Koski-Vähälä (PhD work). Unfortunately, we did not have any field measurements of P exchange between sediment and water available in Yläneenjoki or in any other representative river. (See Lines 304-307).

We redraw figures 5 and 6 in color to make them more illustrative and easier to read.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

After additions and modifications, the article is sufficiently informative for publications. I have no further comments on the content. In some newly added parts of the text, minor editing adjustments need to be made, especially on lines 265, 432, 433 and 504.