the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Jul 2024
Earth observation reveals reduced winter wheat growth and the importance of soil water storing capacity during drought
Abstract. Drought poses increasing challenges to global food production. Knowledge about the influence of drought on crop development and about the role of soil properties in drought risk analysis and mitigating drought impacts at the landscape level is important to guide climate change adaptation. Satellite earth observations can provide area-wide insights into crop growth processes that may help identify risk factors and quantify vulnerability to drought. Here, we evaluate the potential of Sentinel-2 to reveal interactions of plant-growth and soil parameters during variable weather conditions. As a case study, we assess winter wheat growth on 13 fields belonging to commercial farmers in southern Sweden in a dry year and a year with normal weather conditions. To track crop growth, green leaf area index (GLAI) was estimated from satellite imagery using a radiative transfer model. Proxies for winter wheat growth rate, peak GLAI, and the timing of peak GLAI were derived from the GLAI development at the single field level.
We then compared the crop growth proxies between the two years and across the fields and related them to measured soil properties. We found a lower growth rate, lower peak GLAI and earlier peak GLAI in the dry year compared to the year with normal weather conditions. An increase in peak GLAI in the dry year was also shown to be related to a higher growth rate, and this was not shown in the year with normal precipitation. Differences in crop development between years were large for some fields but small for other fields: suggesting that soil properties play a role in crop response to drought. We found that fields with a higher amount of plant available water capacity had better crop performance in the dry year and smaller relative differences in growth rate between the two years. The observed lower growth rate, lower peak GLAI, and earlier peak in the dry year compared to the year with normal weather conditions, demonstrate that satellite imagery can be used to quantify plant-soil-weather interactions at scales relevant to commercial farming. Our investigation serves as a first step towards supporting drought risk management, drought adaptation and communication activities on this important topic.
- Preprint
(850 KB) - Metadata XML
-
Supplement
(357 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-1872', Anonymous Referee #1, 11 Sep 2024
General comments:
This study used a remote sensing product to timely assess crop growth at farm scale. Combining with in-situ measured field soil conditions this work evaluated the effects of soil properties and drought on crop growth. The results contribute to understanding and protecting food security under global warming. The manuscript is well motivated and nicely written. However, I believe the structure and content can be largely improved by focusing on the key novel results and keeping the language concise. There are several concerns that the authors should address to enhance the manuscript.
Main concerns:
- Neither soil properties affect plant water available water or soil water holding capacity affect crop drought stress is exactly new finding. I believe the novelty of this study needs to be clearly stated, such as GLAI at farm level can be used as a reliable indicator of crop growth and stress response to guide management.
- Are there yield data as well?
- Heat stress cannot be separated from this setting. Since accumulative temperature is one variable throughout the analysis, I would suggest including in the introduction and expanding in the discussion the impact of heat on wheat growth. Besides, dry years often have high temperatures too, especially that the DMI is associated with both precipitation and temperature.
- Please indicate field management (e.g., irrigation, fertilisation especially concerning manure and slurry, organic or conventional concerning herbicides) and cultivar for all sites in both years in M&M if the information is available. Only mentioning in the discussion that “minimised variation in these factors by selecting fields that were managed by the same farmer in 2018 and in 2021” does not address the uncertainties.
- I am glad that field management is finally brought up although in very end of the discussion (L318ff). I suggest expanding on the management practices especially that the authors highlighted them to contribute to climate change adaptation, considering there are studies on how management and tillage affect plant drought responses concerning crop performance including growth and yield. Although it is emphasised that commercially operated field studies are important, separating from field experiments and controlled potted experiments, the outcome from field research and potted experiments should not be dismissed. Instead, studies focusing on the difference among wheat cultivars (e.g., Tewolde et al. 2006), which is not directly comparable with the results presented in this study, are discussed. Several relevant studies to consider:
- Khan S, Anwar S, Shaobo Y, Gao ZQ, Sun M, Ashraf MY, Ren AX, Yang ZP. 2020. Soil water consumption, water use efficiency and winter wheat production in response to nitrogen fertilizer and tillage. PeerJ 8: e8892.
- Li YB, Hou RX, Tao FL. 2020. Interactive effects of different warming levels and tillage managements on winter wheat growth, physiological processes, grain yield and quality in the North China Plain. Agriculture Ecosystems & Environment 295: 106923.
- Sun, Q., A. K. Gilgen, R. Wittwer, G. von Arx, M. G. A. van der Heijden, V. H. Klaus, and N. Buchmann (2024), Drought effects on trait space of winter wheat are independent of land management, New Phytol, 243(2), 591-606.
- Wittwer, R. A., V. H. Klaus, E. Miranda Oliveira, Q. Sun, Y. Liu, A. K. Gilgen, N. Buchmann, and M. G. A. van der Heijden (2023), Limited capability of organic farming and conservation tillage to enhance agroecosystem resilience to severe drought, Agricultural Systems, 211, 103721.
- Please carefully check the references. I am not sure whether there are mistakes with citation tools or multiple papers were cited for the wrong reasons. For instance, I did not see LAI or GLAI data in Irfan Ullah et al. 2021and Mroz et al. 2023 so that it is unreasonable to state “some studies found no relationships between timing of peak GLAI and crop yield (Irfan Ullah et al. 2021; Mroz et al. 2023) (L282)”. Other instances include “GLAI at the heading stage of winter wheat has been shown to decrease with a high degree of soil compaction (Lipiec et al. 1991) (L56)” - LAI was used in the cited study. Are LAI and GLAI interchangeable? In the referred work, “Soil physical properties and growth of spring barley as related to the degree of compactness of two soils”, spring barley rather than winter wheat was studied.
- In the last paragraph of discussion 4.1, it reads very confusing as the terms are switching between many relevant but distinguished concepts, such as green (aboveground?) biomass as approximated by GLAI and total crop biomass, biomass and yield, growth rate and crop performance, timing of peak GLAI and heading stage. Please also consider clarifying plant available water capacity, soil water storing capacity, and soil water retention and avoiding interchanging the terms.
- Discussing the individual index of growth rate interpreted from GLAI, peak GLAI and timing of peat GLAI in this setting would only make sense after separating from the drought response (e.g., L283ff). You could consider statistical methods such as mixed effects models for such analysis.
Minor points:
- Certain content in the introduction can be rearranged, such as consider moving “The Sentinel satellites provide a high spatial resolution optical imagery of up to 10 metres… (L63ff)”, “Radiative transfer models describe the relationship between leaf and canopy traits and spectral properties of plants using physical principles…” to M&M.
- Please be consistent and use “peak GLAI” rather than only “peak” (e.g., temperature sum at peak GLAI) in text and figure legend.
- In Fig. 2, please explain the dashed line in the caption of legend.
- I wonder if Fig. 5 is essential to keep in the main text. It only serves to show one value. Consider move it to the supplement.
- Statements such as “Due to similar weather conditions (i.e., similar DMI) across all fields within a specific year, the varying crop responses to drought stress between fields imply that additional factors than the weather must have had an impact on crop development (L289ff)” does not bring any new information. Please keep the language concise throughout.
- There is a Tab. S1 and a Table S1 in the supplement. Please sort it out.
- Data availability should include the data used for analysis in this study too. Suggest including GLAI curves with the inferred slopes for all sites in the supplement.
Line-specific comments:
L13ff “Knowledge about …”: This sentence is very long and complicated. Please simplify or separate into shorter sentences.
*L17 “interactions of plant-growth and soil parameters during variable weather conditions…”: might be an overstatement for this study; the potential impact of soil properties on plant drought responses can be interpreted from the analysis, however not the feedback from the plant growth on the soil. Maybe some interactive effects of soil properties and weather conditions on crop growth but not clearly analysed. Besides, I am not sure about “variable weather conditions” by comparing two field seasons that are 3 years apart. Please revise this sentence.
L18 “a dry year and a year with normal weather conditions”: Suggest mentioning these are the year 2018 and year 2021.
L24 “An increase in peak GLAI in the dry year…": seems to be comparing among sites where differences should not be addressed as an increasement. Please revise.
L26 “suggesting that soil properties play a role in crop response to drought”: Before making this statement, differences among fields in e.g., field management, cultivar, and local climate need to be addressed. Same goes to L333ff in conclusion.
L27 “a higher amount of plant available water capacity…”: Should be “higher amount of plant available water” or something like “higher soil capacity for plant available water”; Unclear what “crop performance” refers to.
L28 “observed lower growth rate, lower peak GLAI, and earlier peak…”: suggest changing to “lower estimated growth rate”; please add “GLAI” after “earlier peak”.
L29ff: please add “in” before “the year” for “compared to the year”; “demonstrate that satellite imagery can be used to quantify plant soil-weather interactions at scales relevant to commercial farming” is not well supported by the results and analysis such as there is no “plant soil-weather interactions” nor it is quantified.
L30ff “Our investigation serves as a first step towards supporting drought risk management, drought adaptation and communication activities on this important topic”: I would strongly suggest being specific on what the “first step” is for, e.g., for the GLAI analysis using satellite imagery at field/farm level, than leaving it too broad, that e.g., involving communication activities would be a stretch.
L44ff “Soils with higher resilience to drought…”: What is soil resilience to drought? This sentence is unclear. Please revise.
L57 “there is still limited information about how soil properties affect crop development under various weather conditions”: I would say there has been more than limited studies which should be referred here to summarise the long-researched soil property impact on crop growth, although the results may vary by crop species, scales, and other environmental factors which should not be simply omitted.
L79 “around 40% of the cropping areas with winter wheat in Northern Europe had yields below the 10th percentile” I found this stat not very straightforward as it is not intuitive for what percentile 40% of the cropping areas should yield below. Suggest revising and improve clarity.
L122ff “The GLAI was derived from the radiative transfer model PROSAIL following the approach described in Graf et al. (2023)… We randomly generated combinations of leaf and canopy parameters according to a uniform or Gaussian distribution (Table S1)”: Table S1 (Tab. S1 in supplement) is identical to Table 5 in Graf et al. 2023 using the same parameters beside GLAI. I am not familiar with this approach, but I wonder if Graf et al. 2023 and two other papers cited by Graf et al. 2023 (Wocher et al. 2020 and Danner et al. 2021) on the parameters in the table should be cited here as well.
- Wocher M., Berger K., Danner M., Mauser W., Hank T. RTM-based dynamic absorption integrals for the retrieval of biochemical vegetation traits Int. J. Appl. Earth Obs. Geoinf., 93 (2020), Article 102219, 10.1016/j.jag.2020.102219
- Danner M., Berger K., Wocher M., Mauser W., Hank T. Efficient RTM-based training of machine learning regression algorithms to quantify biophysical & biochemical traits of agricultural crops ISPRS J. Photogramm. Remote Sens., 173 (2021), pp. 278-296, 10.1016/j.isprsjprs.2021.01.017
L200 Fig. 3 (c): I am confused by this figure that it seems the number of fields presented here don’t sum up to 13.
L249 Fig. 6 (c): Apologies for the clumsy question - as there are 5 circles in the plots, do they correspond to 10, 20, 30, 40, and 50 or 55? Why is the 5th circle having a smaller gap than the other ones?
L268 “Reduced wheat biomass during drought has been shown in earlier studies (Villegas et al. 2001; Zhang et al. 2018), and according to Villegas et al. (2001), the decrease in biomass und droughtis mainly due to lower growth rate”: please correct this sentence.
L273ff “highlighting the importance of faster growth to mitigate drought impacts”: this statement can only be drawn given specific drought severity, duration, and timing which I found missing in the discussion.
L283 “not unambiguous”: please revise.
Fig. S1: Suggest including long-term climate variability e.g., shown as error bar.
Citation: https://doi.org/10.5194/egusphere-2024-1872-RC1 -
AC1: 'Reply on RC1', Hanna Sjulgård, 14 Nov 2024
Dear reviewer,
Thank you for reading the manuscript and for providing valuable feedback and constructive comments that will help to improve our manuscript. In the following, we address each of your comments:
Neither soil properties affect plant water available water or soil water holding capacity affect crop drought stress is exactly new finding. I believe the novelty of this study needs to be clearly stated, such as GLAI at farm level can be used as a reliable indicator of crop growth and stress response to guide management.
Thank you for this comment. The novelty of our work is that we demonstrate that we can use satellite images and these methods of estimated growth proxies to identify drought impacts on crops, and that we even see the impact of soil properties, at the scale relevant for farmers in their fields. We will state the novelty of the study in the revised manuscript more clearly at the end of the introduction, in the discussion and in the conclusions.
Are there yield data as well?
This is an on-farm study, with the challenge that not all farms have yield data available, so unfortunately we only have yield data for some of the fields. However, in the PhD thesis of the first author (Sjulgård 2024), a strong correlation was found between the peak GLAI and winter wheat yield in the year 2021 when including the fields with yield data available and with additional farm fields in the same region.
Sjulgård H. 2024. The potential of agricultural management to alleviate extreme weather impacts on Swedish crop production. Acta Universitatis Agriculturae Sueciae. (2024:84). https://doi.org/10.54612/a.107ri2j3pt
Heat stress cannot be separated from this setting. Since accumulative temperature is one variable throughout the analysis, I would suggest including in the introduction and expanding in the discussion the impact of heat on wheat growth. Besides, dry years often have high temperatures too, especially that the DMI is associated with both precipitation and temperature.
The year 2018 was unusually dry and warm in Sweden, and we focused on the drought because the lack of precipitation has been referred to as the main reason for the yield losses (Bakke et al. 2020, Beillouin et al. 2020, SMHI 2018). Beillouin et al. 2020 stated about cereal yields in 2018 that “In Northern Europe, no negative impacts of high temperature were observed, possibly because maximum temperature values averaged over JJA rarely exceed 25°C” (where JJA refers to June, July and August). However, even though the maximum temperatures were not very extremely high for a longer period of time, there could have been some days with higher temperatures. However, heat stress mostly affect winter wheat negatively during grain formation which was not assessed in our study, and growth is not as much influenced by heat, where the optimum temperature for wheat growth is as high as 20° to 25°C (FAO). Therefore, we did not include heat stress in this study. Regarding the DMI, because higher temperatures potentially can cause higher evapotranspiration and in turn more drought stress, we include both temperature and precipitation data in our aridity index DMI.
Bakke SJ, Ionita M, Tallaksen LM. 2020. The 2018 northern European hydrological drought and its drivers in a historical perspective. Hydrol Earth Syst Sci. 24(11):5621–5653. https://doi.org/10.5194/hess-24-5621-2020
Beillouin D, Schauberger B, Bastos A, Ciais P, Makowski D. 2020. Impact of extreme weather conditions on European crop production in 2018. Philos Trans R Soc Lond B Biol Sci. 375(1810):20190510. https://doi.org/10.1098/rstb.2019.0510
SMHI. 2018. Skörd av spannmål, trindsäd och oljeväxter 2018 [Internet]. [accessed 2024 Oct 25]. https://jordbruksverket.se/om-jordbruksverket/jordbruksverkets-officiella-statistik/jordbruksverkets-statistikrapporter/statistik/2022-03-02-skord-av-spannmal-trindsad-och-oljevaxter-2018--preliminara-uppgifter-for-riket
FAO. Wheat growth and physiology - E. Acevedo, P. Silva, H. Silva [Internet]. [accessed 2024 Nov 13]. https://www.fao.org/4/y4011e/y4011e06.htm
Please indicate field management (e.g., irrigation, fertilisation especially concerning manure and slurry, organic or conventional concerning herbicides) and cultivar for all sites in both years in M&M if the information is available. Only mentioning in the discussion that “minimised variation in these factors by selecting fields that were managed by the same farmer in 2018 and in 2021” does not address the uncertainties.
We know that all the fields have conventional cropping systems (and not organic) and this information will be added to the M&M section. We also know which crop was cultivated each year for all fields but do not have information about the cultivar. Lastly, none of the fields was irrigated. However, we will shift the focus of the study to better highlight its key novelty, i.e., the potential of satellite images to assess growth proxies to identify climatic stresses in farm fields. When doing this, the differences in management methods between fields becomes less important.
Of course, management practices could affect the growth and contribute to different growth between fields, but we will not compare exact growth values between fields and make conclusions based on that. Instead, we want to see if the drought stress can be identified with the methods within each field (as significant differences between the years or not). For those comparisons between the years for the same field, as in Fig. 3c, we will continue to include the percentage difference in the growth proxies, as we stated that there are probably small differences in management practices between the years for cultivating the same crop a few years apart for the same farmer. But of course, there could have been differences in management between years, which might have had an effect that we could not capture. In addition, when investigating the relationships to soil properties, we do the comparisons between fields and not within the same field, and as management practices also can influence the soil properties there is still some uncertainty to this. However, we assume that the farmers manage their fields according to best practice.
I am glad that field management is finally brought up although in very end of the discussion (L318ff). I suggest expanding on the management practices especially that the authors highlighted them to contribute to climate change adaptation, considering there are studies on how management and tillage affect plant drought responses concerning crop performance including growth and yield. Although it is emphasised that commercially operated field studies are important, separating from field experiments and controlled potted experiments, the outcome from field research and potted experiments should not be dismissed. Instead, studies focusing on the difference among wheat cultivars (e.g., Tewolde et al. 2006), which is not directly comparable with the results presented in this study, are discussed. Several relevant studies to consider:
Khan S, Anwar S, Shaobo Y, Gao ZQ, Sun M, Ashraf MY, Ren AX, Yang ZP. 2020. Soil water consumption, water use efficiency and winter wheat production in response to nitrogen fertilizer and tillage. PeerJ 8: e8892.
Li YB, Hou RX, Tao FL. 2020. Interactive effects of different warming levels and tillage managements on winter wheat growth, physiological processes, grain yield and quality in the North China Plain. Agriculture Ecosystems & Environment 295: 106923.
Sun, Q., A. K. Gilgen, R. Wittwer, G. von Arx, M. G. A. van der Heijden, V. H. Klaus, and N. Buchmann (2024), Drought effects on trait space of winter wheat are independent of land management, New Phytol, 243(2), 591-606.
Wittwer, R. A., V. H. Klaus, E. Miranda Oliveira, Q. Sun, Y. Liu, A. K. Gilgen, N. Buchmann, and M. G. A. van der Heijden (2023), Limited capability of organic farming and conservation tillage to enhance agroecosystem resilience to severe drought, Agricultural Systems, 211, 103721.
Thank you for the comment. We do not dismiss pot, greenhouse or field plot studies, they are important to understand certain factors, and hence complement on-farm studies like ours. Because most earlier studies which assessed relationships between GLAI and soil properties have been conducted in field experiments, we stated that on-farm research, such as our study, is also important “since the heterogeneity of environmental factors in the landscape is more complex than what can be investigated in typical field plot experiments.” (L60-61). But we will add in the Introduction section a sentence stating that field and potted experiments also is beneficial to make this clear. At the end of the discussion section, we will also add a few sentences about the advantages and disadvantages/challenges of on-farm studies, focusing on our study.
To further add some discussion about the influence of management practices, we will state that management can influence growth and yield, and here the suggested references will be useful.
Please carefully check the references. I am not sure whether there are mistakes with citation tools or multiple papers were cited for the wrong reasons. For instance, I did not see LAI or GLAI data in Irfan Ullah et al. 2021and Mroz et al. 2023 so that it is unreasonable to state “some studies found no relationships between timing of peak GLAI and crop yield (Irfan Ullah et al. 2021; Mroz et al. 2023) (L282)”. Other instances include “GLAI at the heading stage of winter wheat has been shown to decrease with a high degree of soil compaction (Lipiec et al. 1991) (L56)” - LAI was used in the cited study. Are LAI and GLAI interchangeable? In the referred work, “Soil physical properties and growth of spring barley as related to the degree of compactness of two soils”, spring barley rather than winter wheat was studied.
Thanks for carefully reading our manuscript and for noticing this, there was indeed a mistake in the sentence, and it should have been that Irfan Ullah et al. 2021 and Mroz et al. 2023 found no relationships between the timing of heading to wheat yield, and not between the timing of peak GLAI to wheat yield. The timing of the heading stage is estimated around the peak GLAI. However, because that is not always the case we will exclude the sentences in lines 282 and 283 where we mention other studies assessing relationships between the timing of heading and wheat yield.
Thank you also for pointing out a mistake in the reference to Lipiec et al. 1991, who used spring barley in their study. This was correctly referred to in the discussion (“Lipiec et al. (1991) that found decreasing GLAI at the heading stage of spring barley with a higher degree of soil compaction.”; Lines 307-308), but a mistake was made in the introduction where we wrongly wrote winter wheat. We will correct this in the introduction. Green leaf area index is referred to both as GLAI or as LAI in many studies, but some studies also include yellow and dead leaves in the LAI measurements. Lipiec et al. 1991 do not state if they only used green leaves or not, but until the start of senescence, the different leaf area index measurements should be very similar.
In the last paragraph of discussion 4.1, it reads very confusing as the terms are switching between many relevant but distinguished concepts, such as green (aboveground?) biomass as approximated by GLAI and total crop biomass, biomass and yield, growth rate and crop performance, timing of peak GLAI and heading stage. Please also consider clarifying plant available water capacity, soil water storing capacity, and soil water retention and avoiding interchanging the terms.
Thank you for this helpful comment. We will rewrite the text and use the same wordings throughout the manuscript in the revised version.
Discussing the individual index of growth rate interpreted from GLAI, peak GLAI and timing of peak GLAI in this setting would only make sense after separating from the drought response (e.g., L283ff). You could consider statistical methods such as mixed effects models for such analysis.
Thank you for the suggestion. To do so, we will perform multiple linear regressions which include also the average monthly summer DMI for the corresponding year (2018 or 2021) as an explanatory variable.
Minor points:
Certain content in the introduction can be rearranged, such as consider moving “The Sentinel satellites provide a high spatial resolution optical imagery of up to 10 metres… (L63ff)”, “Radiative transfer models describe the relationship between leaf and canopy traits and spectral properties of plants using physical principles…” to M&M.
We will address this suggestion in the revised manuscript and exclude “The Sentinel satellites provide a high spatial resolution optical imagery of up to 10 metres… (L63ff)” from the introduction, because we already mention the revisit time for our study area in the M&M. “Radiative transfer models describe the relationship between leaf and canopy traits and spectral properties of plants using physical principles…” we will keep in the introduction as we think it is good for the reader to already in the introduction understand what radiative transfer models are, because we already there mentioned the use of them. However, we will connect it better to the previous sentence in the revised version of the manuscript.
Please be consistent and use “peak GLAI” rather than only “peak” (e.g., temperature sum at peak GLAI) in text and figure legend.
Thank you for pointing this out, we will change to a consistent wording.
In Fig. 2, please explain the dashed line in the caption of legend.
The dashed line shows the plateau from the linear plateau model. We will explain it in the figure caption in the revised manuscript.
I wonder if Fig. 5 is essential to keep in the main text. It only serves to show one value. Consider move it to the supplement.
Thank you for the suggestion, we agree that it would fit better in the supplements. As suggested earlier, we will also make it into a table with multiple linear regression results instead.
Statements such as “Due to similar weather conditions (i.e., similar DMI) across all fields within a specific year, the varying crop responses to drought stress between fields imply that additional factors than the weather must have had an impact on crop development (L289ff)” does not bring any new information. Please keep the language concise throughout.
We will streamline the text of the whole manuscript and write it as concisely as possible.
There is a Tab. S1 and a Table S1 in the supplement. Please sort it out.
Thank you for pointing that out, we will correct it and rename the tables in the supplements in the revised manuscript.
Data availability should include the data used for analysis in this study too. Suggest including GLAI curves with the inferred slopes for all sites in the supplement.
We will create GLAI development curves in both years for all fields and add into the supplements. We will also make the data of all those graphs available, and all the calculated growth proxies.
Line-specific comments:
L13ff “Knowledge about …”: This sentence is very long and complicated. Please simplify or separate into shorter sentences.
Thank you for pointing that out, we will simplify this sentence in the revised manuscript to “Knowledge about the influence of drought on crop development and about the role of soil properties is important in drought risk analysis and for mitigating drought impacts at the landscape level.”
*L17 “interactions of plant-growth and soil parameters during variable weather conditions…”: might be an overstatement for this study; the potential impact of soil properties on plant drought responses can be interpreted from the analysis, however not the feedback from the plant growth on the soil. Maybe some interactive effects of soil properties and weather conditions on crop growth but not clearly analysed. Besides, I am not sure about “variable weather conditions” by comparing two field seasons that are 3 years apart. Please revise this sentence.
We will rewrite the abstract to better state the novelty of the study in the revised manuscript and then this sentence “interactions of plant-growth and soil parameters...” will not be included.
L18 “a dry year and a year with normal weather conditions”: Suggest mentioning these are the year 2018 and year 2021.
We agree, and we will add the years 2018 and 2021 in the text.
L24 “An increase in peak GLAI in the dry year…": seems to be comparing among sites where differences should not be addressed as an increasement. Please revise.
You are correct, we will change the word “increase” to “higher”.
L26 “suggesting that soil properties play a role in crop response to drought”: Before making this statement, differences among fields in e.g., field management, cultivar, and local climate need to be addressed. Same goes to L333ff in conclusion.
We investigated the differences in local climate, and the monthly average DMI showed that the climatic conditions were similar between fields, as shown in Figure S1. Our results showed that there was a higher winter wheat growth rate during the extremely dry year with higher plant available water capacity of the soils, and less differences in growth rate between the two years with higher plant available water capacity. Due to these relationships, we think it is fair to “suggest that soil properties play a role in crop response to drought”, and we assume that the farmers manage their fields according to best practice. But of course, there could have been differences in management between years, which might have had an effect that we could not capture, which gives some uncertainty to this.
L27 “a higher amount of plant available water capacity…”: Should be “higher amount of plant available water” or something like “higher soil capacity for plant available water”; Unclear what “crop performance” refers to.
Thank you for pointing that out, we will rewrite this sentence to “a higher plant available water capacity”. The word “better crop performance” referred to that we found a faster growth rate, and we will rewrite this in the revised version.
L28 “observed lower growth rate, lower peak GLAI, and earlier peak…”: suggest changing to “lower estimated growth rate”; please add “GLAI” after “earlier peak”.
Thank you for pointing that out; we will correct it in the revised version of the manuscript.
L29ff: please add “in” before “the year” for “compared to the year”; “demonstrate that satellite imagery can be used to quantify plant soil-weather interactions at scales relevant to commercial farming” is not well supported by the results and analysis such as there is no “plant soil-weather interactions” nor it is quantified.
Thank you, we will add “in” before “the year”. We will also rewrite the sentence in the revised version of the manuscript, and instead write that “demonstrate that satellite imagery can be used to discover drought stress responses and relationships between soil properties and plant development at scales relevant to commercial farming.”
L30ff “Our investigation serves as a first step towards supporting drought risk management, drought adaptation and communication activities on this important topic”: I would strongly suggest being specific on what the “first step” is for, e.g., for the GLAI analysis using satellite imagery at field/farm level, than leaving it too broad, that e.g., involving communication activities would be a stretch.
Thank you for the suggestion, with the better focus of the aim of the manuscript (as described in detail above) we will rewrite the abstract in the revised version and remove this statement, because as you say, we do not discuss communication activities in this manuscript.
L44ff “Soils with higher resilience to drought…”: What is soil resilience to drought? This sentence is unclear. Please revise.
We referred to soils which are more drought-tolerant. We will rewrite this sentence to make it more clear. In the revised manuscript, we will change the sentence to “Soils which are more drought-tolerant…”.
L57 “there is still limited information about how soil properties affect crop development under various weather conditions”: I would say there has been more than limited studies which should be referred here to summarise the long-researched soil property impact on crop growth, although the results may vary by crop species, scales, and other environmental factors which should not be simply omitted.
We apologize for the wording, the sentence should have been more detailed mentioning winter wheat GLAI development and extreme weather conditions at the field level (as many studies are conducted in field trials). We will change the sentence to “there is still limited information about how soil properties affect winter wheat GLAI development under extreme weather conditions, at scales relevant to commercial agriculture”.
L79 “around 40% of the cropping areas with winter wheat in Northern Europe had yields below the 10th percentile” I found this stat not very straightforward as it is not intuitive for what percentile 40% of the cropping areas should yield below. Suggest revising and improve clarity.
Instead of explaining the drought in 2018 based on yield losses in line 79 in the Introduction section, we will only explain the dry conditions during year 2018 in the M&M section based on the weather data. This sentence will therefore be removed.
L122ff “The GLAI was derived from the radiative transfer model PROSAIL following the approach described in Graf et al. (2023)… We randomly generated combinations of leaf and canopy parameters according to a uniform or Gaussian distribution (Table S1)”: Table S1 (Tab. S1 in supplement) is identical to Table 5 in Graf et al. 2023 using the same parameters beside GLAI. I am not familiar with this approach, but I wonder if Graf et al. 2023 and two other papers cited by Graf et al. 2023 (Wocher et al. 2020 and Danner et al. 2021) on the parameters in the table should be cited here as well.
- Wocher M., Berger K., Danner M., Mauser W., Hank T. RTM-based dynamic absorption integrals for the retrieval of biochemical vegetation traits Int. J. Appl. Earth Obs. Geoinf., 93 (2020), Article 102219, 10.1016/j.jag.2020.102219
- Danner M., Berger K., Wocher M., Mauser W., Hank T. Efficient RTM-based training of machine learning regression algorithms to quantify biophysical & biochemical traits of agricultural crops ISPRS J. Photogramm. Remote Sens., 173 (2021), pp. 278-296, 10.1016/j.isprsjprs.2021.01.017
Thank you for this observation. Graf et al. (2023) indeed used the work by Wocher et al. (2020) and Danner et al. (2021) as a starting point and updated the GLAI, leaf chlorophyll and leaf carotenoid content using empirical relationships derived from multi-year field phenotyping data. In Graf et al. (2023), where table S1 can be found, it is made clear which contributions came from which study. To make it easier for readers of this manuscript we agree that citing Wocher et al. (2021) and Danner et al. (2021) is appropriate.
L200 Fig. 3 (c): I am confused by this figure that it seems the number of fields presented here don’t sum up to 13.
Thank you for observing this, for unknown reasons the number in almost all bars was one larger than it should have been (e.g., showing 4 instead of 3), but in the text the correct numbers were mentioned. We will correct the figure in the revised version of the manuscript.
L249 Fig. 6 (c): Apologies for the clumsy question - as there are 5 circles in the plots, do they correspond to 10, 20, 30, 40, and 50 or 55? Why is the 5th circle having a smaller gap than the other ones?
The last circle is only the outside border of the plot area and does not belong to any explained variation. The rings are 10%, 20%, 30% and 40% explained variation starting from the smallest.
L268 “Reduced wheat biomass during drought has been shown in earlier studies (Villegas et al. 2001; Zhang et al. 2018), and according to Villegas et al. (2001), the decrease in biomass und drought is mainly due to lower growth rate”: please correct this sentence.
Thank you for pointing that out; we will correct the spelling mistake of “und” to “under” in the revised version of the manuscript.
L273ff “highlighting the importance of faster growth to mitigate drought impacts”: this statement can only be drawn given specific drought severity, duration, and timing which I found missing in the discussion.
Thank you for pointing this out, the statement is a bit too general. We found a relationship between a faster growth rate to peak GLAI in the dry year, and this suggests that a higher growth rate is favourable during dry conditions to obtain higher yield. A higher peak GLAI has been related to higher yield in other studies e.g. by Lambert et al. 2018, He et al. 2020, Yamamoto et al. 2023 and Sjulgård 2024. We will rewrite the sentence in the revised manuscript to “suggesting that a faster growth is important to obtain higher yield during dry conditions.”
Sjulgård H. 2024. The potential of agricultural management to alleviate extreme weather impacts on Swedish crop production. Acta Universitatis Agriculturae Sueciae. (2024:84). https://doi.org/10.54612/a.107ri2j3pt
He J, Shi Y, Zhao J, Yu Z, He J, Shi Y, Zhao J, Yu Z. 2020. Strip rotary tillage with subsoiling increases winter wheat yield by alleviating leaf senescence and increasing grain filling. Crop Journal. 8(2):327–340.
Lambert M-J, Traoré PCS, Blaes X, Baret P, Defourny P. 2018. Estimating smallholder crops production at village level from Sentinel-2 time series in Mali’s cotton belt. Remote Sensing of Environment. 216:647–657. https://doi.org/10.1016/j.rse.2018.06.036
Yamamoto S, Hashimoto N, Homma K. 2023. Evaluation of LAI Dynamics by Using Plant Canopy Analyzer and Its Relationship to Yield Variation of Soybean in Farmer Field. Agriculture. 13(3):609. https://doi.org/10.3390/agriculture13030609
L283 “not unambiguous”: please revise.
As mentioned in an earlier answer, we will exclude the sentences in line 282 and 283 where we mention other studies assessing relationships between the timing of heading and wheat yield.
Fig. S1: Suggest including long-term climate variability e.g., shown as error bar.
Thank you for this suggestion, we will include error bars in this figure in the revised version.
Citation: https://doi.org/10.5194/egusphere-2024-1872-AC1
-
RC2: 'Comment on egusphere-2024-1872', Anonymous Referee #2, 20 Sep 2024
This study shows how Sentinal-2 signals respond to winter wheat growth during drought under different soil types. The paper touches a very interesting and important topic which could add a new insight to the contribution of satellite products for field level purposes. Although the study starts with a nice story and introduction and a very logical database to address this particular issue, but it lacks a lot of critical information and data when it comes to presents and use them for interpretation. This made the interpretation of the results difficult and to some level impossible. Therefore, in my opinion, the paper must be improved substantially by providing more levels of supporting information. There are several concerns that the authors should address to enhance the manuscript.
- What is the novelty of this research? The fact that the different responses of two fields to drought could be related to soil properties is not novel. The author should make it clear how these different signals could be attributed to soil at larger scale and how this analysis lead us to landscape level. Or make the novelty of the analysis very clear.
- The authors showed only two extreme fields (Fig. 3) to present the responses of different fields. However, it is still not clear how this varies in other fields. I strongly recommend that the author provides the results for all 13 fields in the supplementary with information of soil data on each subplot. Without this, we cannot understand how the results of field which had different responses are related to soil.
- Another important lack is related to measurements of soil field capacity. The authors have measured field capacity at -10 KPa for all fields. However, based on Table S1, the clay content varies between 10 and 58%. Measuring FC at -10 KPs is valid for sandy soils and not for clayey soils which is measured at -33 KPa. So this has resulted in a wrong procedure for determining the plant available water capacity that biases the results and conclusions.
- The paper lacks a lot when it comes to presenting management data. It is not clear to what level these variations could be related to management factors. When was planting dates in different years and fields? Are there different cultivars? If they are different, how they differ in terms of phenological development? What are fertilizations? What are disease effects? All these information must be clearly provided and at the end reflected in the interpretation of results. Without this, it was difficult to me to make interpretation.
- The response of crop to drought depends on time of drought and the time of temperature increase. At which phenological stage did drought happen? I think temperature increase play a critical role here. I think here more detailed information on the stage of growth and onset of drought and temperature increase are required. The paper provides zero information on climatic variation between two years. Were the severity of drought and temperature increase the same in all years?
- The authors used the concept of temperature sum. Do they mean growing degree days? If yes, what were the base temperature?
- 3a. The crop response are significantly different even in April. This cannot be related to drought. I think the author should clearly justify using climate data that the drought already started in April. Otherwise these difference is related to factors which was overseen.
- In line 23, the author mentioned that “we found a lower growth rate, lower peak GLAI and earlier peak GLAI”. This is surprising to me, because the fig.3 shows that the peak occurs at about the same time.
- The author showed different vegetation responses in different years, however it is still not clear to what level it is related to yield data. How much these two responses did influence the final yield?
- 5 is confusing. Why the correlation was shown for only growth rate and peak GLA? What about the other?
- Line 314,”in soil properties changing over time”. This is surprising to me, soil properties do not change over 2-3 years which makes such interpretation less convincing.
Citation: https://doi.org/10.5194/egusphere-2024-1872-RC2 -
AC2: 'Reply on RC2', Hanna Sjulgård, 14 Nov 2024
Dear reviewer,
Thank you for reading the manuscript and for your valuable feedback and constructive comments. In the following we have addressed each of your comments:
What is the novelty of this research? The fact that the different responses of two fields to drought could be related to soil properties is not novel. The author should make it clear how these different signals could be attributed to soil at larger scale and how this analysis lead us to landscape level. Or make the novelty of the analysis very clear.
Thank you for this suggestion. Reviewer 1 had a similar comment. The novelty of our work is that we show that we can use satellite images and these methods of estimated growth proxies to identify drought impacts on crops and that we even see the impact of soil properties, at the scale relevant for farmers in their fields. We will more clearly state the novelty of the study in the revised manuscript at the end of the introduction, in the discussion and in the conclusions.
The authors showed only two extreme fields (Fig. 3) to present the responses of different fields. However, it is still not clear how this varies in other fields. I strongly recommend that the author provides the results for all 13 fields in the supplementary with information of soil data on each subplot. Without this, we cannot understand how the results of field which had different responses are related to soil.
Thank you for this suggestion. We do show data for all fields (Fig 3c, Fig 4) and we used Figs 3a and 3b as illustrative examples. However, we will in the revised manuscript also include plots for all 13 fields in the Supplementary materials.
Another important lack is related to measurements of soil field capacity. The authors have measured field capacity at -10 KPa for all fields. However, based on Table S1, the clay content varies between 10 and 58%. Measuring FC at -10 KPs is valid for sandy soils and not for clayey soils which is measured at -33 KPa. So this has resulted in a wrong procedure for determining the plant available water capacity that biases the results and conclusions.
We understand how you are thinking, but within the same study, it becomes difficult to compare the results between soil samples if we use different matric potentials for different soils. Because we have a gradient in texture it also becomes unprecise to divide all soils into sand or clay soils. In Europe, it is common to use -10 kPa as we did and to define plant available water as the difference between the water contents at -10 kPa and wilting point. In e.g. the US the -33 kPa is more commonly used for field capacity (e.g. Nemes et al. 2011).
Nemes A, Pachepsky YA, Timlin DJ. 2011. Toward Improving Global Estimates of Field Soil Water Capacity. Soil Science Society of America Journal. 75(3):807–812. https://doi.org/10.2136/sssaj2010.0251
The paper lacks a lot when it comes to presenting management data. It is not clear to what level these variations could be related to management factors. When was planting dates in different years and fields? Are there different cultivars? If they are different, how they differ in terms of phenological development? What are fertilizations? What are disease effects? All these information must be clearly provided and at the end reflected in the interpretation of results. Without this, it was difficult to me to make interpretation.
The reviewer 1 raised a similar point. The exact sowing date is unknown but note that winter wheat was used in this study, which is sown within a short time window around the middle of September, is almost dormant during the cold winter, and then it starts to grow again in spring. The exact sowing date is less important for autumn-sown crops in comparison to spring-sown crops when comparing the growth during spring due to the winter period, as the start of the growing season in spring can be more important for early growth than the sowing date.
We know that all the fields have conventional cropping systems (and not organic) and this information will be added to the M&M section. We also know which crop was cultivated each year for all fields but do not have information about the cultivar. Lastly, none of the fields was irrigated. However, we will shift the focus of the study to better highlight its key novelty, i.e., the potential of satellite images to assess growth proxies to identify climatic stresses in farm fields. When doing this, the differences in management methods between fields become less important.
Of course, management practices could affect the growth and contribute to different growth between fields, but we will not compare exact growth values between fields and make conclusions based on that. Instead, we want to see if the drought stress can be identified with the methods within each field (as significant differences between the years or not). For those comparisons between the years for the same field, as in Fig. 3c, we will continue to include the percentage difference in the growth proxies, as we stated that there are probably small differences in management practices between the years for cultivating the same crop a few years apart for the same farmer. But of course, there could have been differences in management between years, which might have had an effect that we could not capture. In addition, when investigating the relationships to soil properties, we do the comparisons between fields and not within the same field, and as management practices also can influence the soil properties there is still some uncertainty to this. However, we assume that the farmers manage their fields according to best practices.
The response of crop to drought depends on time of drought and the time of temperature increase. At which phenological stage did drought happen? I think here more detailed information on the stage of growth and onset of drought and temperature increase are required:
Thank you for this suggestion. The unusually dry conditions started already in May in 2018, during the vegetative growth period. We will include this information in the M&M section. In the revised version of the manuscript we will also add a more detailed figure showing the temporal development in average temperature and precipitation during the growing season for the two years.
The paper provides zero information on climatic variation between two years.:
Maybe you accidentally missed the figures in the Supplementary Information. In Fig. S1 we show the average monthly De Martonne Aridity Index between May to July for the two years 2018 and 2021, and also the long-term average between 1991 and 2020 for the fields. See the previous answer about also adding an extra figure.
“Were the severity of drought and temperature increase the same in all years?”:
We included two years in this study, and the point in choosing year 2018 and 2021 was to compare an extremely dry year (2018) with a year with normal weather conditions (2021). Here, Fig S1 shows the differences in average aridity (May to July) between the years.
The authors used the concept of temperature sum. Do they mean growing degree days? If yes, what were the base temperature?
Temperature sum and growing degree days are quite similar, but we used temperature sum because we had average temperature data available for the fields, while growing degree days require maximum and minimum temperatures. The temperature sum was assessed by adding up the daily mean temperatures exceeding the threshold value of 0 °C. As the base temperature, in line 135 we stated that in the calculations of temperature sum, we used “the daily mean temperatures exceeding the threshold value of 0 °C, where growth for winter wheat starts (Porter and Gawith 1999)”.
3a. The crop response are significantly different even in April. This cannot be related to drought. I think the author should clearly justify using climate data that the drought already started in April. Otherwise these difference is related to factors which was overseen.
We understand that Figure 3a and b could confuse readers because they show GLAI plotted against calendar dates, but the analyses were done based on temperature sum. The relationships between the curves therefore look a bit different when plotted using calendar dates or temperature sum. When the curves are plotted against the temperature sum, the start of the curves is more similar. To not confuse the readers, we will plot Fig. 3a and b against temperature sum instead of calendar dates as the new Fig. 3a and 3b. We think it is still important to include some figure with GLAI plotted against the calendar date to see around what time of the year e.g. the peak GLAI occur. We will therefore also include the current Fig. 3a and 3b with GLAI plotted against the calendar dates in the supplementary materials of the revised version of the manuscript.
In line 23, the author mentioned that “we found a lower growth rate, lower peak GLAI and earlier peak GLAI”. This is surprising to me, because the fig.3 shows that the peak occurs at about the same time.
When including all fields in the comparison, there was a significantly earlier peak GLAI in the dry year, as seen in Figure 3c and in Figure 4. For the visual interpretation of Figure 3a and b, we refer to the answer in the previous question.
The author showed different vegetation responses in different years, however it is still not clear to what level it is related to yield data. How much these two responses did influence the final yield?
Reviewer 1 had a similar comment. This is an on-farm study, with the challenge that not all farms provided yield data, so unfortunately we only have yield data for some of the fields. However, in the PhD thesis of the first author (Sjulgård 2024), a strong correlation was found between the peak GLAI and winter wheat yield in year 2021 when including the fields with yield data available and with additional farm fields in the same region. Similarly, the peak GLAI has been related to crop yield in earlier studies e.g. by Lambert et al. 2018, He et al. 2020 and Yamamoto et al. 2023.
Sjulgård H. 2024. The potential of agricultural management to alleviate extreme weather impacts on Swedish crop production. Acta Universitatis Agriculturae Sueciae. (2024:84). https://doi.org/10.54612/a.107ri2j3pt
He J, Shi Y, Zhao J, Yu Z, He J, Shi Y, Zhao J, Yu Z. 2020. Strip rotary tillage with subsoiling increases winter wheat yield by alleviating leaf senescence and increasing grain filling. Crop Journal. 8(2):327–340.
Lambert M-J, Traoré PCS, Blaes X, Baret P, Defourny P. 2018. Estimating smallholder crops production at village level from Sentinel-2 time series in Mali’s cotton belt. Remote Sensing of Environment. 216:647–657. https://doi.org/10.1016/j.rse.2018.06.036
Yamamoto S, Hashimoto N, Homma K. 2023. Evaluation of LAI Dynamics by Using Plant Canopy Analyzer and Its Relationship to Yield Variation of Soybean in Farmer Field. Agriculture. 13(3):609. https://doi.org/10.3390/agriculture13030609
5 is confusing. Why the correlation was shown for only growth rate and peak GLA? What about the other?
With the 5 we assume you refer to Figure 5, but in Figure 5 we show the results for all growth proxies, and the timing of peak GLAI (“temperature sum at peak”) is also part of the figure. However, based on comments from reviewer #1 we will remove the figure and present the results of the relationships between the growth proxies in a supplementary table instead in the revised version of the manuscript.
Line 314,”in soil properties changing over time”. This is surprising to me, soil properties do not change over 2-3 years which makes such interpretation less convincing.
We kindly disagree with the reviewer. Many soil properties are dynamic and may change within weeks, months, years or decades, see e.g.:
Berhe AA. 2019. Chapter 3 - Drivers of soil change. In: Busse M, Giardina CP, Morris DM, Page-Dumroese DS, editors. Developments in Soil Science. Vol. 36. Elsevier; p. 27–42. https://doi.org/10.1016/B978-0-444-63998-1.00003-3
Tan C, Cao X, Yuan S, Wang W, Feng Y, Qiao B. 2015. Effects of Long-term Conservation Tillage on Soil Nutrients in Sloping Fields in Regions Characterized by Water and Wind Erosion. Sci Rep. 5(1):17592. https://doi.org/10.1038/srep17592
Mohan D, Mamrutha HM, Khobra R, Singh G, Singh GP. 2022. Relevance of height, heading and maturity in productivity enhancement of wheat. INDIAN JOURNAL OF GENETICS AND PLANT BREEDING. 82(01):31–37. https://doi.org/10.31742/IJGPB.82.1.5
Due to the impact of changes in management practises on soil properties, we stated in line 315-318 that “a number of studies has shown only small year-to-year changes in soil organic carbon content (Krauss et al. 2020), water content at field capacity (Alam et al. 2014) and bulk density (Alam et al. 2014; Alnaimy et al. 2020) within given soil management systems.“. Because the same crop is grown in both 2018 and 2021 and the same farmer is managing the fields between years, we assume that the management practises were similar, and that the soil properties that were measured in this study would probably not have changed significantly from year 2018 to 2021.
Citation: https://doi.org/10.5194/egusphere-2024-1872-AC2
-
AC2: 'Reply on RC2', Hanna Sjulgård, 14 Nov 2024
Dear reviewer,
Thank you for reading the manuscript and for your valuable feedback and constructive comments. In the following we have addressed each of your comments:
What is the novelty of this research? The fact that the different responses of two fields to drought could be related to soil properties is not novel. The author should make it clear how these different signals could be attributed to soil at larger scale and how this analysis lead us to landscape level. Or make the novelty of the analysis very clear.
Thank you for this suggestion. Reviewer 1 had a similar comment. The novelty of our work is that we show that we can use satellite images and these methods of estimated growth proxies to identify drought impacts on crops and that we even see the impact of soil properties, at the scale relevant for farmers in their fields. We will more clearly state the novelty of the study in the revised manuscript at the end of the introduction, in the discussion and in the conclusions.
The authors showed only two extreme fields (Fig. 3) to present the responses of different fields. However, it is still not clear how this varies in other fields. I strongly recommend that the author provides the results for all 13 fields in the supplementary with information of soil data on each subplot. Without this, we cannot understand how the results of field which had different responses are related to soil.
Thank you for this suggestion. We do show data for all fields (Fig 3c, Fig 4) and we used Figs 3a and 3b as illustrative examples. However, we will in the revised manuscript also include plots for all 13 fields in the Supplementary materials.
Another important lack is related to measurements of soil field capacity. The authors have measured field capacity at -10 KPa for all fields. However, based on Table S1, the clay content varies between 10 and 58%. Measuring FC at -10 KPs is valid for sandy soils and not for clayey soils which is measured at -33 KPa. So this has resulted in a wrong procedure for determining the plant available water capacity that biases the results and conclusions.
We understand how you are thinking, but within the same study, it becomes difficult to compare the results between soil samples if we use different matric potentials for different soils. Because we have a gradient in texture it also becomes unprecise to divide all soils into sand or clay soils. In Europe, it is common to use -10 kPa as we did and to define plant available water as the difference between the water contents at -10 kPa and wilting point. In e.g. the US the -33 kPa is more commonly used for field capacity (e.g. Nemes et al. 2011).
Nemes A, Pachepsky YA, Timlin DJ. 2011. Toward Improving Global Estimates of Field Soil Water Capacity. Soil Science Society of America Journal. 75(3):807–812. https://doi.org/10.2136/sssaj2010.0251
The paper lacks a lot when it comes to presenting management data. It is not clear to what level these variations could be related to management factors. When was planting dates in different years and fields? Are there different cultivars? If they are different, how they differ in terms of phenological development? What are fertilizations? What are disease effects? All these information must be clearly provided and at the end reflected in the interpretation of results. Without this, it was difficult to me to make interpretation.
The reviewer 1 raised a similar point. The exact sowing date is unknown but note that winter wheat was used in this study, which is sown within a short time window around the middle of September, is almost dormant during the cold winter, and then it starts to grow again in spring. The exact sowing date is less important for autumn-sown crops in comparison to spring-sown crops when comparing the growth during spring due to the winter period, as the start of the growing season in spring can be more important for early growth than the sowing date.
We know that all the fields have conventional cropping systems (and not organic) and this information will be added to the M&M section. We also know which crop was cultivated each year for all fields but do not have information about the cultivar. Lastly, none of the fields was irrigated. However, we will shift the focus of the study to better highlight its key novelty, i.e., the potential of satellite images to assess growth proxies to identify climatic stresses in farm fields. When doing this, the differences in management methods between fields become less important.
Of course, management practices could affect the growth and contribute to different growth between fields, but we will not compare exact growth values between fields and make conclusions based on that. Instead, we want to see if the drought stress can be identified with the methods within each field (as significant differences between the years or not). For those comparisons between the years for the same field, as in Fig. 3c, we will continue to include the percentage difference in the growth proxies, as we stated that there are probably small differences in management practices between the years for cultivating the same crop a few years apart for the same farmer. But of course, there could have been differences in management between years, which might have had an effect that we could not capture. In addition, when investigating the relationships to soil properties, we do the comparisons between fields and not within the same field, and as management practices also can influence the soil properties there is still some uncertainty to this. However, we assume that the farmers manage their fields according to best practices.
The response of crop to drought depends on time of drought and the time of temperature increase. At which phenological stage did drought happen? I think here more detailed information on the stage of growth and onset of drought and temperature increase are required:
Thank you for this suggestion. The unusually dry conditions started already in May in 2018, during the vegetative growth period. We will include this information in the M&M section. In the revised version of the manuscript we will also add a more detailed figure showing the temporal development in average temperature and precipitation during the growing season for the two years.
The paper provides zero information on climatic variation between two years.:
Maybe you accidentally missed the figures in the Supplementary Information. In Fig. S1 we show the average monthly De Martonne Aridity Index between May to July for the two years 2018 and 2021, and also the long-term average between 1991 and 2020 for the fields. See the previous answer about also adding an extra figure.
“Were the severity of drought and temperature increase the same in all years?”:
We included two years in this study, and the point in choosing year 2018 and 2021 was to compare an extremely dry year (2018) with a year with normal weather conditions (2021). Here, Fig S1 shows the differences in average aridity (May to July) between the years.
The authors used the concept of temperature sum. Do they mean growing degree days? If yes, what were the base temperature?
Temperature sum and growing degree days are quite similar, but we used temperature sum because we had average temperature data available for the fields, while growing degree days require maximum and minimum temperatures. The temperature sum was assessed by adding up the daily mean temperatures exceeding the threshold value of 0 °C. As the base temperature, in line 135 we stated that in the calculations of temperature sum, we used “the daily mean temperatures exceeding the threshold value of 0 °C, where growth for winter wheat starts (Porter and Gawith 1999)”.
3a. The crop response are significantly different even in April. This cannot be related to drought. I think the author should clearly justify using climate data that the drought already started in April. Otherwise these difference is related to factors which was overseen.
We understand that Figure 3a and b could confuse readers because they show GLAI plotted against calendar dates, but the analyses were done based on temperature sum. The relationships between the curves therefore look a bit different when plotted using calendar dates or temperature sum. When the curves are plotted against the temperature sum, the start of the curves is more similar. To not confuse the readers, we will plot Fig. 3a and b against temperature sum instead of calendar dates as the new Fig. 3a and 3b. We think it is still important to include some figure with GLAI plotted against the calendar date to see around what time of the year e.g. the peak GLAI occur. We will therefore also include the current Fig. 3a and 3b with GLAI plotted against the calendar dates in the supplementary materials of the revised version of the manuscript.
In line 23, the author mentioned that “we found a lower growth rate, lower peak GLAI and earlier peak GLAI”. This is surprising to me, because the fig.3 shows that the peak occurs at about the same time.
When including all fields in the comparison, there was a significantly earlier peak GLAI in the dry year, as seen in Figure 3c and in Figure 4. For the visual interpretation of Figure 3a and b, we refer to the answer in the previous question.
The author showed different vegetation responses in different years, however it is still not clear to what level it is related to yield data. How much these two responses did influence the final yield?
Reviewer 1 had a similar comment. This is an on-farm study, with the challenge that not all farms provided yield data, so unfortunately we only have yield data for some of the fields. However, in the PhD thesis of the first author (Sjulgård 2024), a strong correlation was found between the peak GLAI and winter wheat yield in year 2021 when including the fields with yield data available and with additional farm fields in the same region. Similarly, the peak GLAI has been related to crop yield in earlier studies e.g. by Lambert et al. 2018, He et al. 2020 and Yamamoto et al. 2023.
Sjulgård H. 2024. The potential of agricultural management to alleviate extreme weather impacts on Swedish crop production. Acta Universitatis Agriculturae Sueciae. (2024:84). https://doi.org/10.54612/a.107ri2j3pt
He J, Shi Y, Zhao J, Yu Z, He J, Shi Y, Zhao J, Yu Z. 2020. Strip rotary tillage with subsoiling increases winter wheat yield by alleviating leaf senescence and increasing grain filling. Crop Journal. 8(2):327–340.
Lambert M-J, Traoré PCS, Blaes X, Baret P, Defourny P. 2018. Estimating smallholder crops production at village level from Sentinel-2 time series in Mali’s cotton belt. Remote Sensing of Environment. 216:647–657. https://doi.org/10.1016/j.rse.2018.06.036
Yamamoto S, Hashimoto N, Homma K. 2023. Evaluation of LAI Dynamics by Using Plant Canopy Analyzer and Its Relationship to Yield Variation of Soybean in Farmer Field. Agriculture. 13(3):609. https://doi.org/10.3390/agriculture13030609
5 is confusing. Why the correlation was shown for only growth rate and peak GLA? What about the other?
With the 5 we assume you refer to Figure 5, but in Figure 5 we show the results for all growth proxies, and the timing of peak GLAI (“temperature sum at peak”) is also part of the figure. However, based on comments from reviewer #1 we will remove the figure and present the results of the relationships between the growth proxies in a supplementary table instead in the revised version of the manuscript.
Line 314,”in soil properties changing over time”. This is surprising to me, soil properties do not change over 2-3 years which makes such interpretation less convincing.
We kindly disagree with the reviewer. Many soil properties are dynamic and may change within weeks, months, years or decades, see e.g.:
Berhe AA. 2019. Chapter 3 - Drivers of soil change. In: Busse M, Giardina CP, Morris DM, Page-Dumroese DS, editors. Developments in Soil Science. Vol. 36. Elsevier; p. 27–42. https://doi.org/10.1016/B978-0-444-63998-1.00003-3
Tan C, Cao X, Yuan S, Wang W, Feng Y, Qiao B. 2015. Effects of Long-term Conservation Tillage on Soil Nutrients in Sloping Fields in Regions Characterized by Water and Wind Erosion. Sci Rep. 5(1):17592. https://doi.org/10.1038/srep17592
Mohan D, Mamrutha HM, Khobra R, Singh G, Singh GP. 2022. Relevance of height, heading and maturity in productivity enhancement of wheat. INDIAN JOURNAL OF GENETICS AND PLANT BREEDING. 82(01):31–37. https://doi.org/10.31742/IJGPB.82.1.5
Due to the impact of changes in management practises on soil properties, we stated in line 315-318 that “a number of studies has shown only small year-to-year changes in soil organic carbon content (Krauss et al. 2020), water content at field capacity (Alam et al. 2014) and bulk density (Alam et al. 2014; Alnaimy et al. 2020) within given soil management systems.“. Because the same crop is grown in both 2018 and 2021 and the same farmer is managing the fields between years, we assume that the management practises were similar, and that the soil properties that were measured in this study would probably not have changed significantly from year 2018 to 2021.
Citation: https://doi.org/10.5194/egusphere-2024-1872-AC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 408 | 89 | 31 | 528 | 157 | 11 | 17 |
- HTML: 408
- PDF: 89
- XML: 31
- Total: 528
- Supplement: 157
- BibTeX: 11
- EndNote: 17
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1