Evaluating Irrigation Efficiency with Performance Indicators: A Case Study of Citrus in the East of Spain

Round 1

Reviewer 1 Report

Agronomy-911362-peer-review-v1

I very much appreciate this paper and the attempts by the authors to consider irrigation efficiency and productivity via the establishment of indicators.  This is the kind of paper that fits well with the journal Agronomy.  And it contains important attempts to consider how agricultural water management is to be improved.  And it reads well (though some improvements to the English language are needed) and it is structured into some main sections.   

However the scale of revisions is significant.  I think it should be published once major revisions are attended to or it could be rejected and accepted again as a new paper.  My comments on the paper are divided into three main categories; English language; substantive comments and more minor comments.  I hope it is clear from the below why major revisions are needed.

1. There is a need for the English language, grammar and syntax to be improved

The authors need to have their paper reviewed and improved by a professional copy-editor to ensure the English language is correct.   There are many small errors in the paper and here are some of them:

E.g. ‘ben’ rather than ‘been’ on line 27.

Some sentences simply don’t make sense: “ The effect of point excess irrigation is also studied with the previous factors on the indicators.” (Line 23 in the abstract).  What is point excess? 

The title does not make sense “Evaluating the Irrigation Efficiency with Performance Indicators: A Case of Study Citrus in the East of Spain”.  Perhaps it should be “Evaluating Irrigation Efficiency with Performance Indicators: A Case Study of Citrus in the East of Spain”.  ?

Use “an” rather than “the” in “Finally, the artificial intelligence system is used to predict productive efficiency” (Line 24)

Line 41 change “food supply in sustainable manners” to “food supply in a sustainable manner”

No need to make ‘climate change’ a proper noun by capitalising the first letters.

Line 173 edit “bestprocedures”

Line 350 grammar needs correcting – it should be “However, according to this general analysis,” or However, according to these general analyses,”

The use of the word “majoritarian” is a little arcane – better to use ‘major’.

And there are many other examples to be corrected (which is not the responsibility of this reviewer). 

2. Substantive points

My main criticism and I am sorry to say this is that the paper spends a lot of space discussing variables that are not really central to the topic of the paper as set out in the abstract.   My reading of the abstract is that the paper should be about irrigation management, and irrigation and productive efficiency as understood via some indicators.  Here is the relevant sentence from the abstract: “ In this paper we propose and test different indicators for service delivery performance, productive efficiency and economic efficiency.”   I would therefore make the paper much clearer about the derivation of the indicators and then show how they reflect irrigation management. 

Therefore I urge the authors to drop some variables and their associated discussion that are not really that important to the topic of ‘how the citrus farmers managed their irrigation and how we know how they did that”.  For example Figures 4 to 7 take up two whole pages, but I can’t see any really interesting relationships or trends for the purpose of explaining irrigation management and why their chosen indicators are so worthwhile and significant in revealing irrigation management.  For example why would the professional or non-professional status of the farmer tell us something about irrigation management?  One might hypothesise that their status has a bearing on irrigation management, but that hypothesis would be open to criticism since it would be only one of many factors that influence whether a farmer is a good or poor irrigator.  And actually the paper does not explain this relationship in sufficient depth for the status to be confirmed and therefore germane to their abstract. 

And, much more significantly, how are the variables that appear to define irrigation practices actually tied to good or poor performance?  For example in lines 447-450 does the application of more water mean the irrigators are more precise and accurate (good irrigators) or are simply over-irrigating (poor irrigators).  Here is the relevant text: “ farmers are applying higher irrigation than NP farmers, and they are classified as two different groups. In average, P farmers are using 543.76 l/m2 while NP farmers irrigate with 219 l/m2 in their plots. Thus, according to the irrigation, P farmers are using a higher amount of water than NP farmers (147% of water).”  But there is no further information given as to whether 542.76 l/m2 is ‘efficient, adequate, accurate’.  Thus the volume of water applies appears to be an indicator (implicitly implied by the text), but the reader can’t see how it indicates something.

To continue my argument about variables that should be dropped – how is it possible that ‘YEAR’ defines good and poor irrigation?   Surely the amount of water applied is not a function of time, but is a function of climate or a function of other political-economic factors (such as a declining price of water over time)?  But this reasoning is not explained, so the current version of the paper does in fact seem to be saying that what defines irrigation performance is simply whether one is sitting in 2014 or 2017. 

With only five years of data, are the authors sure that they have employed the right methodology?  They seem to be looking for trends of variables that explain the amount of water applied from a inferential statistical point of view.  I would expect instead the data to be used in a different way to fully explore the variation in irrigation behaviours and how their chosen indicators show this behaviour.  But I would do this in a way that makes technical sense and in ways that mean the indicators are revealing something.  For example, relative water supply could be used to show there might be greater variability of water applied in a drier year as compared to a wet year.  Or that amongst a group of say 50 irrigators, the variation is too high to be recorded as ‘good system management’.  Or if one farmer gets consistently good yields, then from the water data, we can see she or he ‘accurate’ and is not over or under irrigating.  Or if a farmer is consistently under-irrigating, then what might explain that practice? 

In other words, the writing team must begin with their key points and then build the data and data-analysis around that.  If they don’t do this, the paper says it will do one thing but actually spends many pages doing another thing. 

I therefore wonder if the research team have gone about this research using an approach of what is ‘doable’ with the collected data.  In other words, taking Figures 5-7, we see what are deemed to be independent variables connected to land productivity.  The fact that the collected data can be analysed to show this, does not mean it needs to go into a paper about water productivity tied to water management revealed by indicators.    

I do not understand the need to have two tables of variable and data types; Table 1 and Table 2.  I don’t see the clear difference and these tables do not make clear which is primary data, or modelled data or recalculated data.  I also cannot see which Table 1 or 2 references the data from the modelled CROPWAT.   I would advise ONE table of data, with columns explaining the variable’s label, unit, source. 

The methodology of the approach is to correlate outputs of fruit production (different variables) and connect them to a number of management inputs using available management data.  The authors need to make clear that they did not measure any inputs or outputs and that everything is based on secondary obtained data or modelled data.

It is not clear in the current version if the irrigation applied volumes are either modelled based on agro-met data or are actuals based on farmers metered records. 

It is not clear if the variables are applied to single farmers or to ‘groups of farmers’ or to ‘between farmer’ analyses.  In other words, is relative water supply a function of a changing water supply to one farmer or the changes in water supply arising with supplies to different farmers. 

Related to this, the authors need to be crystal clear that they are introducing errors related to modelled agromet data from a met station that is not local to the production area (the authors do acknowledge this but it is hidden – see lines 253-254).  The ensuing discussion should not be spuriously accurate since ultimately the researchers do not know how accurate the farmers are in irrigating efficiently.

I don’t understand equation 5.  The unit should be ‘depth of water applied’ (m3/m2) and not volume? 

I don’t understand equation 6?  If this is about income over cost, why are the units kg/m2 and m3/m2?

Also early on in the paper, I think the paper’s authors need to keep closely to the topic of water management and efficiency.  I suggest the authors are more strict with how they introduce topics that are not central to their argument.  I would therefore take out references to factors that while appear to be related, are not strongly connected to the paper’s main thesis.  E.g. line 44, reference is made to networking.  I don’t see the relevance of this?   And the authors don’t then refer to networking?

The authors need to stick with the terms they introduce and define.  It seems that ‘economic efficiency’ later on becomes ‘economic balance’. 

My understanding is that ‘qualitative’ data relates to social interview data, and not to independent variables such as crop type or farmer type.   As far as I can see the paper has no qualitative data resulting from farmers opinions.

The paper is frustrating to read for other reasons.  E.g. here is an example from line 343-345: “On the other hand, the data of applied irrigation (see Table 4) shows a clear pattern; the irrigation increases year by year. From 2013 to 2017, the amount of applied irrigation has increased in 87.27 l/m2, which represents 44.2% of the initial irrigation.”  But 44.2% is less than half (meaning the irrigators are applying 44% of the initial irrigation?) so how does the percentage apply? 

How does the mix of citrus species and varieties change so dramatically in only four years?  Perennial trees take a long time to grow and cannot be changed on an annual basis, so I don’t understand this kind of sentence in lines 381-382 “Oranges represent 75.93%, 66.09%, 60%, and 49.18% in the following years. On the other side, the cropping of tangerines has been extended.”

This paragraph (lines 484-492) is indicative of the paper’s inability to really comment on what is going on.  Despite all the work, the authors end up speculating about irrigation management during the period under research plus the reader is not guided by the authors as to what the resulting indicators mean.  What does an economic indicator of 18 actually signify in terms of ‘excellent, good, poor performance’ of irrigation and crop management?    

The real discussion on the indicators in section 4.4.1 is somewhat difficult to digest.  The same applies to 4.4.2.  This is where the meat of the paper should be, but I feel the authors are discussing a statistical analysis of the indicators rather than how the indicators reveal something about irrigation management.  Put another way, I don’t think readers of this paper end up knowing that ‘citrus irrigation in East Spain is exceptionally well management at the system and orchard scale as shown by xx and xx indicators”.  Instead I end up trying to navigate the statistical variances of five indicators applied to things like ‘year’, ‘variety’, ‘extra irrigation’, plot size’ and so on.  I struggle to put together the story of irrigation in this irrigation system.  And many variables are omitted e.g. the age of the drip equipment, its maintenance, its design fitted to local conditions. 

Section 4.4.3 contains some really interesting discussion but it hinges on or refers to Figure 14 which I can’t see makes much sense.  The x axes mix category variables of farmer status and variety and this makes no sense in terms of what the graphs mean. 

I am not equipped to comment on the ANN section, but my view is that this section is not central to the paper and in a future version could be dropped.  It appears to be another way of expressing some interesting statistical patterns but are not tied to real water management. 

Line 636 to 646 contains a method which potentially is interesting – asking what is behind an individual’s performance?  But the paragraph does not explain a farmer’s high performance in a way that is satisfactory.  Surely it cannot be what the paragraph implies which is that the high performance is a function solely of plot size and professional status.  There must be other irrigation and farming practices that explain consistently high agronomic and productive performance.  What is an ‘alternate bearing’? 

The discussion section 5 is interesting and some of these considerations should guide the paper.  However small things frustrate the reader – when referring to ‘productivity’ do the authors mean ‘water productivity’ or ‘land productivity’? 

In line 675, the word ‘consume’ is used, whereas in fact the irrigation records are probably ‘water applied’. 

The section “5.3. Are the Professional farmers more efficient in water management than Non-Professional farmers?” returns to the point I made above.  It is a worthwhile question to ask, but I hope the authors see that they are not in a position with their current data and indicators to answer that question with any great credibility.  There is so much one would need to analyse to get to a definitive answer – especially if other factors such as remoteness from an agromet station is a driving factor in making incorrect irrigation scheduling choices. 

I appreciate that the researchers cannot go back and re-run their research, but the paper needs to be ‘water management-facing’ rather than ‘statistics-facing’.  Also in my view Section 5 and Section 4 should be merged into ‘results and discussion’.  In summary, my view is that the paper should be restructured to take each indicator by itself, and then use a sub-section to explain how it reveals something about citrus irrigation.  Each sub-section should get its story right as to how the indicator is derived, what the indicator means, how it was applied to the field data and what the results derived signify about irrigation management in the period 2014-2017. 

In any re-write, it might help to make it very clear that the research paper is about ‘post-hoc’ analysis of benchmarking-type indicators rather than of diagnostic day-to-day water scheduling and management indicators.  This distinction might help remove the concern of this reviewer that the paper’s indicators are supposed help guide farmers manage their irrigation (I don’t think so), rather I think the indicators are better suited to looking at broad trends of performance. 

Furthermore, the authors, despite being honest about their short-comings of their research, need to pin down what has worked for them and what has not.  Lines 667-683 is honest, but the text introduces doubt about the validity of the whole paper.   The recommendation to farmers to locate a nearer met station also introduces doubt – as it raises questions such as; “how do the farmers currently know when to irrigate and which fields need irrigating”.  This question should be central within this paper so that the indicators help the farmers make those decisions or for a third party to gauge whether their methods of irrigation scheduling are ‘accurate’ (or not). 

Mindful of all my comments, I hope the authors are not dissuaded from resubmitting either as a revised paper or new paper.  This kind of analysis is really important and my wish is to see something in the literature that irrigation managers can refer to. 

Other or more minor points

My preference is to express land area as hectares (not as m2) and irrigation applied as a ‘depth equivalent’ (mm) not as litres/m2.  Part of the problem is that the units in the paper change from litres / m2, to m3/m2 and m2 to m3 part of the problem is that without being ‘eagle-eyed’ one has to read very carefully the units.  Line 348 is a good example of how the current version trips up the unwary reader “Regarding productivity, the higher average value in kg/m2 was found in 2013 with 2.59kg/m2;”.  Do the authors see that in a paper about water productivity, this discussion is actually about land productivity?  And the only way to know that is the use of m2 not m3?   A small difference which for a social scientists or lay person might make for quite a frustrating read. 

I urgently suggest that the words ‘plot’ and ‘plots’ are clarified.  Do the authors mean plots are ‘orchards’ or statistical plots as in graphs?  The paper uses the words in different ways which makes it very difficult to follow the discussion.  Once clarified, do not mix.  

The symbols in the equations need to be explained.  For example, what is ‘m’?  Or ‘n’ or ‘u’?

Line 349 – why the spurious accuracy?  “197.50l/m2”

Please help reader understand scales ‘enterprise, district, plot’ etc.  E.g. by adding indicative sizes in km2 or hectares.  Is plot (orchard) scale 0.5 or 1 hectare or 10 hectares?

In lines 80-84 please clarify and help the reader understand whether each indicator studies one aspect of citrus farming, or does one indicator study three aspects of citrus farming. 

It is normal to write numbers in full when below 10, and above 10 then to use the numeric form.  So ‘3’ should be ‘three’.

Line 215 what is ‘three status’?

Line 300-301 – is the type of computer really relevant in a paper published in 2020?

I would expect to see the Acknowledgements also thank the anonymous reviewers – even though we bring a long review like this. 

Author Response

Please see our reply in the attached file.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors propose an interesting comparison of indicators for service delivery performance, productivity and economic efficiency to be used in the analysis of citrus orchards. These indexes can help farmers to identify better irrigation methods for their plots.

The article presents extensive description of methods and results, and the MS quality is good: the structure is well arranged, the presentation is clear and the case study is suitable for its application, although, as the authors themselves say, better accuracy could be achieved with a greater amount of data, especially in the use of ANN model.

Overall, I have some minor suggestions to further improve the MS quality:

• English language and style are fine but there are some minor typos/spell-check in the text that a careful editing could identify and correct;
• Please check the unit of measurement in equations 4, 5 and 6;
• I think that in Table 3, three decimal places are excessive. The precision could be the same used in the rest of the text;
• Figures quality overall is very good, but I think Figures from 7 to 10 could be sharper to match the quality of other images.

Author Response

Please see our reply in the attached file.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

I'm very impressed with how the authors have responded to my comments.  As a result of their thorough work, I am pleased to say that the paper can go forward.