Geomatics and EO Data to Support Wildlife Diseases Assessment at Landscape Level: A Pilot Experience to Map Infectious Keratoconjunctivitis in Chamois and Phenological Trends in Aosta Valley (NW Italy)

Round 1

Reviewer 1 Report

The authors well address all previous comments by this reviewer.

 

Author Response

Response to Reviewer 1 Comments

 

We would like to thank reviewers for their appropriate comments and helpful suggestions that have been carefully considered. Majority of provided suggestions highlighted gaps in the text and were really useful to improve, we hope, paper quality. In blue referees can find their comments, in red authors’ actions to reply/satisfy requests.

In particular, the synthesis of reviewers’ comments suggested a deep revision in paper organization and harmonization. Consequently, you will find some structural changes aimed at simplifying paper reading and make content more effective.

All comments were carefully evaluated and for the most of them corrections and integrations have been provided.

 

Point 1: The authors well address all previous comments by this reviewer.

 

Response 1: We would like to thank the reviewer for his/her comment. A general English comprehensive editing has been done taking into account also the suggestion proposed by others referee. In order to avoid to report here all the changes performed, we kindly suggest the reviewer to see the revised paper.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

This manuscript has been much improved and the results are interesting. The new figures add value to the text. However, the current version is still hard to follow in places due to a large number of grammatical errors throughout the text.

The introduction is still too long and lacks focus. Some of the text in this section might be better added to the discussion. The discussion section is rather short and could be improved by adding some of the statements and references from the latter half of the introduction.

Figure 1 doesn't add any value to the manuscript.

The revised manuscript should be edited and proof read for grammatical errors

Author Response

Response to Reviewer 2 Comments

 

We would like to thank reviewers for their appropriate comments and helpful suggestions that have been carefully considered. Majority of provided suggestions highlighted gaps in the text and were really useful to improve, we hope, paper quality.

In particular, the synthesis of reviewers’ comments suggested a deep revision in paper organization and harmonization. Consequently, you will find some structural changes aimed at simplifying paper reading and make content more effective.

All comments were carefully evaluated and for the most of them corrections and integrations have been provided.

 

 

Point 1: This manuscript has been much improved and the results are interesting. The new figures add value to the text. However, the current version is still hard to follow in places due to a large number of grammatical errors throughout the text.

 

The introduction is still too long and lacks focus. Some of the text in this section might be better added to the discussion. The discussion section is rather short and could be improved by adding some of the statements and references from the latter half of the introduction.

 

Figure 1 doesn't add any value to the manuscript.

 

The revised manuscript should be edited and proof read for grammatical errors

Response 1: The referee is right. A general comprehensive editing has been done taking into account your useful suggestion. Grammatical errors throughout the text has been correct. We kindly invite the referee to see them on the revised paper.

 

The suggestion given by the reviewer about the introduction has been followed. The introduction has been considerably shortened and some parts have been moved to the discussion section as the referee has wisely reported as follow:

[Introduction] “Geomatics and satellite remote sensing represent a useful analysis tool in several technical-scientific fields [1]. Nowadays remote sensing is widely used in many fields like agronomy, forestry and environment in general. Nevertheless, veterinary and faunistic-related applications are still limited and, often, characterized by an improper use of satellite platforms and EO data, that makes desirable a significant improvement [2].

On the international scene, only few research groups are currently properly exploiting all the potentialities that Geomatics (remote sensing included) and digital geographical data could offer to the veterinary sector. The most of works concern parasitology and virology for etiological and epidemiological studies [3] and diagnosis and medical history [4]. In these context, satellite remote sensing has assumed a great interest in the last years [5]; in fact, Earth Observation (EO) data are, proficiently, used to feed meteorological and climatological models with the aim of generating predictive scenarios of zoonosis spread and outbreaks [6].

Firstly, the most of studies in the veterinary and health sectors used remote sensing to describe environmental conditions; this occurred especially with reference to malaria in Africa and Asia [7; 8].  Presently, epidemiologists are currently adopting remote sensing to investigate a variety of vector-borne diseases. Associations between remote sensing-derived environmental variables (e.g. temperature, humidity, land cover, etc.) and vector density are used to map and characterize vector habitats [9]. The basic idea is that remotely sensed data can contain dynamic predictors of Earth’s processes suitable for describing niche preferences of some medically important host diseases mechanisms. Moreover, because of their continuity of acquisition, remotely sensed data provide a synoptic representation of environment at proper spatial and temporal scales [10].

Meteorological and EO data are jointly used also for diseases analysis. For example, outbreaks of diarrheal disease and, specifically, cholera were analyzed by a new modeling approach based on satellite data to produce cholera risk maps in several regions of globe [8], supporting the idea that the ongoing EO technology transfer is making possible to investigate new patterns in a systemic point of view [11].

Given the veterinary and public health impact of vector-borne diseases, there is a clear and immediate need to map and monitor local landscape attitude to encourage emergence and spread of these diseases. Current approaches for predicting disease risks are mostly neglecting key features of landscape related to the functional habitat of vectors (or hosts) and, hence, of the pathogen [12].

A global satellite-based monitoring of proper climate variables could help to map occurring anomalies with the aim of predicting spatial distribution of risk related to emergence and propagation of disease vectors. Such information could provide sufficient lead-time for outbreak prevention and potentially reduce burden and spread of ecologically coupled diseases.

Additionally, remote sensing could have an important role in the comprehension of patho-system dynamic. With reference to the so called “disease triangle”, including host, pathogens/vectors and environment, remote sensing, and geomatics in general, could support scientists and decision-makers to better understand the role of environmental patterns and, therefore, explore its complex relations with the other parts of patho-system [13; 14]. A good example is represented by atmospheric pollution that was recognized to increase sensitivity to pulmonary diseases, as the last pandemic event, coronavirus (SARS-CoV-2), has suggested [15].

GIS studies about endo- and ecto-parasitoses of veterinary interest, with particular reference to zoonoses agents, represent today the greatest contribution to veterinary and faunistic sectors [16]. For many years the World Organization for Animal Health located in Paris (OIE) and the World Health Organization (WHO) in Geneva have been underlining the importance of Geomatics and remote sensing applications [17] in the "One Health" perspective.

This work, with reference to a regional case study, investigates remote sensing potentialities for describing relationships between environment and diseases affecting wildlife at landscape level. Moreover, it is intended to describe the effects of climate change onto the vegetation component, with special concern about pastures. The study area corresponds to the entire Aosta Valley Region located in the Italian Western Alps. In particular, a new analysis approach is presented to operate at landscape level to analyze if and how environmental factors could condition the occurrence of infectious keratoconjunctivitis (IKC, Mycoplasma conjunctivae) in chamois. IKC is a contagious disease for domestic and wild ruminants (Caprinae and Ovinae) [18]. In chamois, the disease can be serious [19] and, as in other wild ruminants, blindness can occur [20], with consequent death of the animal from trauma (e.g. fall from cliffs or starvation) [21]. The period of mountain pasture is risky for the potential contact between domestic and wild infected animals; along the years, several outbreaks have been reported in wild ungulates in the Alps [22] and this is the reason that makes monitoring / surveillance plans still active. IKC caused by Mycoplasma conjunctivae is a complex disease of domestic and wild Caprinae, with great variations in the clinic-pathological and epidemiological picture. In wildlife, IKC is sometimes associated with high mortality [23; 24]. It has been suggested that the pathogenesis of IKC is influenced by host predispositions, virulence of M. conjunctivae strains, secondary infections, and environmental factors [25]. Sex and age imbalance in affected populations were observed in severe outbreaks [26], indicating that age and social behavior, including sexual segregation, may be important risk factors. Differently, differences in virulence between different strains do not seem to play a major role; mycoplasmal load is obviously associated to the presence and severity of signs. However, the driver of mycoplasmal multiplication in the host is unknown. Environmental factors might have a role, regarding both the expression of the disease in individual cases and the onset of an outbreak in a population [24]. The underlying hypothesis of this work is that remote sensing could support comprehension of the role of environmental patterns in conditioning IKC patho-system, and related pathologies, as for other diseases. Altitude, air quality, and UV light have been discussed as possible predisposing factors for IKC in wild ungulates along with overcrowding [27]. Multiple outbreaks of IKC in Alpine ibex and Alpine chamois populations have been described in literature [28]. It was consistently detected in Pyrenean and Alpine chamois (Rupicapra p. pyrenaica) populations, as well as in sheep flocks (17.0% of sheep) and occasionally in mouflon (Ovis aries musimon) from the Pyrenees (22.2% in one year/area); statistically associated with ocular clinical signs only in chamois. Chamois populations showed different infection dynamics with low but steady prevalence (4.9%) and significant yearly fluctuations (0.0%– 40.0%) between the period 2008-2015. Persistence of specific M. conjunctivae strain clusters in wild host populations is demonstrated for six and nine years. Cross-species transmission between chamois and sheep and chamois and mouflon were also sporadically evidenced. Host population characteristics and M. conjunctivae strains resulted in different epidemiological scenarios in chamois, ranging from the fading out of the mycoplasma to the epidemic and endemic long-term persistence. These findings highlight the capacity of M. conjunctivae to establish diverse interactions and persist in host populations, also with different transmission conditions. Different outbreaks of infectious keratoconjunctivitis (IKC) affecting alpine chamois and ibex in the western and central Swiss Alps and Aosta Valley were recorded in the period 2001-2019 [29].Between the years 2001-2003, in Switzerland, Mycoplasma conjunctivae was identified from conjunctival swabs by means of a nested PCR in 27 of the 28 chamois tested. The outbreaks occurred in an area covering 1590 km². Deep valleys acted as a barrier to the spread of the disease. Many of the affected animals were juveniles, and more females than males died of IKC. The disease was more common during the summer and autumn. The chamois affected by IKC were found at altitudes between 550 and 3200 m. The estimated overall mortality was less than 5 per cent, but more than 20 per cent have probably died locally. In some outbreaks, mortality can reach 30 per cent, as, for example, in chamois in Italy, France and Switzerland, and hundreds of chamois may die. Major outbreaks were recorded in 2001-2003 [30] and 2016-2018 [31]. With these premises, in this work, two types of analysis were performed: one aimed at exploring, by remotely sensed data, phenological metrics (PMs) and evapotraspiration (ET) trends of vegetation; one investigating correlation between PMs and ET versus IKC prevalence. PMs/ET analysis was based on TERRA MODIS image time series ranging from 2000 to 2019. Ground data about IKC were available for a shorter time range: 2009 - 2019. Consequently, PMs and ET trends investigation were done for the whole times range (2000-2019); conversely, correlation analysis was achieved with reference to the 2009-2019 period.” (please see lines 47-182)

 

[Discussions] “The functional roles of domestic and wild host populations in infectious keratoconjunctivitis (IKC) epidemiology have been extensively discussed claiming a domestic reservoir for the more susceptible wild hosts; in the most of cases all deductions were based on limited data.

With the aim to better assess IKC epidemiology in complex host-pathogen alpine systems, the long-term infectious dynamics and molecular epidemiology of Mycoplasma conjunctivae has been investigated in all host populations from different areas in the Pyrenees and Occidental Alps. Overall, independent M. conjunctivae sylvatic and domestic cycles occurred at the wildlife-livestock interface in the alpine ecosystems with sheep and chamois as the key host species for each cycle, and mouflon as a spill-over host. Although outbreaks of IKC have been described in Austria, France, Italy, Slovenia and Switzerland, descriptive studies of the role of environmental patterns and to model the outbreaks on a large scale have often been incomplete, owing to the difficulty of detect fundamentals pattern that can affecting IKC spread in chamois that living in remote, inaccessible mountain regions. Under this scenario the remote sensing techniques and EO data can give certainly a huge hand in the understanding and development of possible forecasting models, as we have tried to do in the present work. With these premises the present study, was intended to explore and propose a method based on free accessible EO data to partially close the above-mentioned knowledge gap. 

In Aosta Valley (NW Italy) PMs and ET (as measured from the above mentioned EO data) proved to significantly changed their values in the last 20 years, with a continuous progressive trend observable for all of them. In terms of strength of changes, an average delay of EOS was observed of about 2.6 days, independently from the altitude class. SOS proved to averagely anticipate of about 2 and 3 days per year at lower (< 2000 m) and higher (> 2000 m) altitudes, respectively. Consequently, LOS is enlarging of about 4.7 and 6.5 days per year at lower (< 2000 m) and higher (> 2000 m) altitudes, respectively. While looking at the entire period (2000-2019) MAXVI proved to be significantly changing, showing a positive variation (about +0.09) at lower altitude and no variations at higher one. This can be explained admitting that at lower altitudes, in Aosta Valley, grasslands and pastures are often irrigated. Consequently, farmers can vary water release regimes to face climate change effects (higher temperatures, in particular) with the result of moving forage yields (that NDVI is a predictor of) to higher values.

Differently, where more natural (not managed) systems are located (higher altitudes) the increasing of yearly MAXVI can be only related to glacier melting that could compensate the increasing of water requirement (as confirmed by the ET analysis) by vegetation: glaciers are, in fact, dramatically reducing in Aosta Valley. Moreover, another compensating action could come from the surrounding forest areas that have been proved to tolerate summer heatwaves.

With reference to ET, a significant increasing trend was observed, independently from altitude.  Eight days water requirement from vegetation appears to averagely increase of about 0.05 Kg·m^-2 (about 0.5%) every year for a total increase of about 1 Kg·m^-2 in 20 years (2000-2019), corresponding to a percentage difference in water requirement from vegetation of about 8%. This could be possibly explained by the increasing of biomass production (well represented by MAXVI) and by the enlargement of the growing season, that, consequently make vegetation needing more water yearly.

As far as PMs/ET and IKC Pr correlation is concerned, some interesting findings came out.

A 2^nd order polynomial model was found to well approximate the most of relationships, making possible to support scenarios generation of IKC spreading for forecasting issues.

ET proved to be a good predictor of IKC Pr, with no significant conditioning by altitude. EOS seems to operate good predictions of Pr at lower altitudes, while LOS at medium-high ones. This probably depends on wildlife dynamics, that in autumn and at the beginning of winter, looks for grassland (food) especially at lower altitudes where, in that period, grass is not covered by snow and is wetter. In the case of SOS, a possible interpretation can rely on the fact that, if vegetative season lasting increase, animals descent from mountain to valley can be postponed, thus inducing a higher probability of interaction between potential guests and sick animals at higher altitudes. This certainly can increase also the probability of exposure of animals to disease.

With reference to ET it is mainly related to Pr at lower altitudes where micro-local temperature and humidity favor vegetation growth and limit soil drought, determining a higher attractiveness for chamois. At higher altitudes ET capability of predicting Pr becomes weaker probably due to a lower availability of biomass and a shorter phenological season.

Authors are conscious that this work just introduces a new way to manage wildlife health problem and cannot be retained conclusive. In the nearer future, more disaggregated investigations should be done, and other areas possibly considered. Nevertheless, the proposed approach is sufficiently innovative in the context of wildlife veterinary and, we hope, could open a new interesting trend to map wildlife diseases and related zoonosis risk associated with the interaction between wild animals and domestic ones. A radical change is expected also by technicians and institutional subjects in their ordinary procedures for recording and manage ground data. In fact, the greatest limit to expand and more focus this research relied on the format of ground data that could be obtained only aggregated at regional level with no information concerning the specific place where each analyzed animal was found. We invite all involved players to carefully consider the possibility of georeference every ground observation that comes to their laboratories. Georeferencing of ground data is at the basis of an effective and reliable spatial based approach like the one here proposed, where EO data (especially if available over a long time span) play a crucial role. Anyway, this work proved that spatially based forecasting models can be reasonably calibrated for generating maps of risk concerning wildlife diseases and zoonosis spreading in a certain area. Moreover, it showed that relationship between IKC and PMs/ET are probably chancing in terms of strength; in fact, we demonstrated that all the considered predictors are suffering from a significant change possibly related to the ongoing climate change. Consequently, we expect that future approaches should more properly rely on contemporary data spatially distributed in place of aggregated data temporally distributed like the one we processed for this work.” (please see lines 607-691)

 

Figure 1 has been mantained despite the observation given by the referee.

this choice was due to two reasons: the first one, because the other referees, conversely, judged it useful to the present work (in any case, respecting the initial figure, changes have been made following the indications of the referees) and the second one, because we try to summarize the potential and possible role of remote sensing as possible tool for a better understanding of the patho-system in the health sector. This aspect is highlighted both in the introduction and in the conclusions.

Author Response File: Author Response.docx

Reviewer 3 Report

The objective of this study is to outline a new approach to incorporate geomatics and remote sensing data into a monitoring program for infectious keratoconjunctivitis in chamois in the Aosta Valley Region in the Italian Western Alps. The authors use environmental data from a variety of Earth observation products, and they calculate changes in phenology in different “favorable” zones identified from altitude and CORINE 2018 land cover types. The authors hypothesize the following: The underlying hypothesis of this work is that remote sensing could support comprehension of the 119 role of environmental patterns in conditioning IKC pathosystem, and related pathologies, as already 120 happen for other diseases. Although the problem is interesting and a nice use of remote sensing data to investigate a problem important to veterinary health, clarity is needed regarding the methods used and their appropriateness to avoid Type I errors in Analysis 2, leading me to recommend a major revision of the manuscript. Additionally, because Remote Sensing is an English language journal, I recommend that an improvement to the grammar and sentence structure be completed prior to publication. Please change “illness triangle” to “disease triangle” Additionally, Figure 1 shows a Venn Diagram of different components of the disease triangle, including Host, Environment, and Pathogen. In the center is disease. Remote sensing is added on the bottom with an overlap on pathogen and conducive host, suggesting that remote sensing can contribute to only these two components. This is not an accurate placement of remote sensing. Remote sensing should be able to help identify habitats for hosts, suitable environments, and their geographic overlap with pathogen systems. The graphs resulting from Analysis 1 have different y-axes between FAV1, FAV2, and FAV3, even though the values area not much different. I recommend plotting these graphs on the same y-axis, so the reader can see whether the trend was steeper for an individual FAV category compared to others for each phenological calculation. Description of Land Use data. Please include a reference to the Google Earth Engine page that provides the description of the data product. The wording used here is nearly exactly the same, warranting a very clear reference: https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_CORINE_V20_100m Lines 410 – 419: The authors state that they used Pearson’s correlation to identify correlations between individual phenological metrics in Analysis 2. Although the authors were careful to decompose the time series analysis in Analysis 1, using the TimeSat software package, they do not appear to have done this for Analysis 2. If the time series is not decomposed, Type I errors can occur because remaining temporal autocorrelation can drive the correlation value. In this case, there is not a true correlation between the two time series, but they share a similar autocorrelation structure. I recommend using the CCF function in R to analyze the time series for Analysis 2 (or a similar time series analysis function from another software program: https://www.rdocumentation.org/packages/tseries/versions/0.1-2/topics/ccf Lines 475 – 493: Several sentences discuss reasons why certain results could be observed and I recommend that they be moved to the discussion section. Lines 513 – same comment, beginning with “EOS behavior at lower altitudes probably depends on wildlife dynamics…” Please check the remainder of this section and move sentences that discuss possible reasons behind observed results to the discussion section. Lines 495-500: The authors state that “model calibration was achieved after removing outliers from data.” The IKC prevalence data is from 2009 – 2019, so only 10 years. How many outliers were removed? Also, the methods section indicated that a Pearson’s correlation was used to identify correlations between IKC prevalence and phenological metrics, but the results report on 2nd order polynomial regression. Which method was used? Table 6. When did Yearly Cumulative ET become a variable being considered? A search of the document does not mention this variable in the methods section. Line 555 – This study has a small number data points to be able to provide robust forecasting. Please revise this sentence.

Author Response

Response to Reviewer 3 Comments

 

We would like to thank reviewers for their appropriate comments and helpful suggestions that have been carefully considered. Majority of provided suggestions highlighted gaps in the text and were really useful to improve, we hope, paper quality.

In particular, the synthesis of reviewers’ comments suggested a deep revision in paper organization and harmonization. Consequently, you will find some structural changes aimed at simplifying paper reading and make content more effective.

All comments were carefully evaluated and for the most of them corrections and integrations have been provided.

 

 

Point 1: The objective of this study is to outline a new approach to incorporate geomatics and remote sensing data into a monitoring program for infectious keratoconjunctivitis in chamois in the Aosta Valley Region in the Italian Western Alps. The authors use environmental data from a variety of Earth observation products, and they calculate changes in phenology in different “favorable” zones identified from altitude and CORINE 2018 land cover types. The authors hypothesize the following: The underlying hypothesis of this work is that remote sensing could support comprehension of the 119 role of environmental patterns in conditioning IKC pathosystem, and related pathologies, as already 120 happen for other diseases. Although the problem is interesting and a nice use of remote sensing data to investigate a problem important to veterinary health, clarity is needed regarding the methods used and their appropriateness to avoid Type I errors in Analysis 2, leading me to recommend a major revision of the manuscript. Additionally, because Remote Sensing is an English language journal, I recommend that an improvement to the grammar and sentence structure be completed prior to publication. Please change “illness triangle” to “disease triangle” Additionally, Figure 1 shows a Venn Diagram of different components of the disease triangle, including Host, Environment, and Pathogen. In the center is disease. Remote sensing is added on the bottom with an overlap on pathogen and conducive host, suggesting that remote sensing can contribute to only these two components. This is not an accurate placement of remote sensing. Remote sensing should be able to help identify habitats for hosts, suitable environments, and their geographic overlap with pathogen systems. The graphs resulting from Analysis 1 have different y-axes between FAV1, FAV2, and FAV3, even though the values area not much different. I recommend plotting these graphs on the same y-axis, so the reader can see whether the trend was steeper for an individual FAV category compared to others for each phenological calculation. Description of Land Use data. Please include a reference to the Google Earth Engine page that provides the description of the data product. The wording used here is nearly exactly the same, warranting a very clear reference: https://developers.google.com/earth-engine/datasets/catalog/COPERNICUS_CORINE_V20_100m Lines 410 – 419: The authors state that they used Pearson’s correlation to identify correlations between individual phenological metrics in Analysis 2. Although the authors were careful to decompose the time series analysis in Analysis 1, using the TimeSat software package, they do not appear to have done this for Analysis 2. If the time series is not decomposed, Type I errors can occur because remaining temporal autocorrelation can drive the correlation value. In this case, there is not a true correlation between the two time series, but they share a similar autocorrelation structure. I recommend using the CCF function in R to analyze the time series for Analysis 2 (or a similar time series analysis function from another software program: https://www.rdocumentation.org/packages/tseries/versions/0.1-2/topics/ccf Lines 475 – 493: Several sentences discuss reasons why certain results could be observed and I recommend that they be moved to the discussion section. Lines 513 – same comment, beginning with “EOS behavior at lower altitudes probably depends on wildlife dynamics…” Please check the remainder of this section and move sentences that discuss possible reasons behind observed results to the discussion section. Lines 495-500: The authors state that “model calibration was achieved after removing outliers from data.” The IKC prevalence data is from 2009 – 2019, so only 10 years. How many outliers were removed? Also, the methods section indicated that a Pearson’s correlation was used to identify correlations between IKC prevalence and phenological metrics, but the results report on 2nd order polynomial regression. Which method was used? Table 6. When did Yearly Cumulative ET become a variable being considered? A search of the document does not mention this variable in the methods section. Line 555 – This study has a small number data points to be able to provide robust forecasting. Please revise this sentence.

Response 1: The referee is right. A general comprehensive editing has been done taking into account his/her useful suggestion. Grammatical errors throughout the text have been corrected. We kindly invite the referee to look at the revised paper.

 

As suggested the illness triangle has been changed in disease triangle (please see lines 86-94).

 

The Venn Diagram of different components of the disease triangle, including host, environment, and pathogen in Figure 1 has been changed putting remote sensing in the middle. We are agreed with the referee and his/her following statement that “…Remote sensing should be able to help identify habitats for hosts, suitable environments, and their geographic overlap with pathogen systems.” (please see lines 86-94).

 

The graphs resulting from Analysis 1 that having different y-axes between FAV1, FAV2, and FAV3, have been changed according to referee suggestion (please see lines 502-509)

Considering the description of Land Use data, as suggested by the referee, it has been included a reference taking into account the Google Earth Engine page that provides the description of the data product (please see line 337).

 

In Analysis 2 we have performed the same analysis done in Analysis 1. We were careful to decompose the time series also in Analysis 2, using the TimeSat software package. In order to avoid possible mis-understanding we have better explained as follow: “After demonstrating that significant trends could be recognized affecting PMs and ET values from MOD13Q and MOD16A2 datasets, respectively, we tested their potential correlation with IKC Pr at regional level. As far as IKC/ET comparison was concerned the Yearly Cumulative ET was calculated. The a priori hypothesis was that a change in environmental conditions could drive to a change in IKC occurrences (Pr value). It is worth to remind that Pr data were supplied aggregated at regional level, while PMs and ET measures were mapped at pixel level over the whole area. Consequently, all deductions refer to general trends that could be possibly improved if more distributed and geolocated data of IKC Pr were available.

IKC Pr was tested against all the computed metrics and modelled by a 2^nd order polynomial regression.” (please see lines 447-460).

Considering about several sentences that discuss reasons in certain results, as suggested by the referee they are moved to the discussion section as follow:

“The functional roles of domestic and wild host populations in infectious keratoconjunctivitis (IKC) epidemiology have been extensively discussed claiming a domestic reservoir for the more susceptible wild hosts; in the most of cases all deductions were based on limited data.

With the aim to better assess IKC epidemiology in complex host-pathogen alpine systems, the long-term infectious dynamics and molecular epidemiology of Mycoplasma conjunctivae has been investigated in all host populations from different areas in the Pyrenees and Occidental Alps. Overall, independent M. conjunctivae sylvatic and domestic cycles occurred at the wildlife-livestock interface in the alpine ecosystems with sheep and chamois as the key host species for each cycle, and mouflon as a spill-over host. Although outbreaks of IKC have been described in Austria, France, Italy, Slovenia and Switzerland, descriptive studies of the role of environmental patterns and to model the outbreaks on a large scale have often been incomplete, owing to the difficulty of detect fundamentals pattern that can affecting IKC spread in chamois that living in remote, inaccessible mountain regions. Under this scenario the remote sensing techniques and EO data can give certainly a huge hand in the understanding and development of possible forecasting models, as we have tried to do in the present work. With these premises the present study, was intended to explore and propose a method based on free accessible EO data to partially close the above mentioned knowledge gap. 

In Aosta Valley (NW Italy) PMs and ET (as measured from the above mentioned EO data) proved to significantly changed their values in the last 20 years, with a continuous progressive trend observable for all of them. In terms of strength of changes, an average delay of EOS was observed of about 2.6 days, independently from the altitude class. SOS proved to averagely anticipate of about 2 and 3 days per year at lower (< 2000 m) and higher (> 2000 m) altitudes, respectively. Consequently, LOS is enlarging of about 4.7 and 6.5 days per year at lower (< 2000 m) and higher (> 2000 m) altitudes, respectively. While looking at the entire period (2000-2019) MAXVI proved to be significantly changing, showing a positive variation (about +0.09) at lower altitude and no variations at higher one. This can be explained admitting that at lower altitudes, in Aosta Valley, grasslands and pastures are often irrigated. Consequently, farmers can vary water release regimes to face climate change effects (higher temperatures, in particular) with the result of moving forage yields (that NDVI is a predictor of) to higher values.

Differently, where more natural (not managed) systems are located (higher altitudes) the increasing of yearly MAXVI can be only related to glacier melting that could compensate the increasing of water requirement (as confirmed by the ET analysis) by vegetation: glaciers are, in fact, dramatically reducing in Aosta Valley. Moreover, another compensating action could come from the surrounding forest areas that have been proved to tolerate summer heatwaves.

With reference to ET, a significant increasing trend was observed, independently from altitude.  Eight days water requirement from vegetation appears to averagely increase of about 0.05 Kg·m^-2 (about 0.5%) every year for a total increase of about 1 Kg·m^-2 in 20 years (2000-2019), corresponding to a percentage difference in water requirement from vegetation of about 8%. This could be possibly explained by the increasing of biomass production (well represented by MAXVI) and by the enlargement of the growing season, that, consequently make vegetation needing more water yearly.

As far as PMs/ET and IKC Pr correlation is concerned, some interesting findings came out.

A 2^nd order polynomial model was found to well approximate the most of relationships, making possible to support scenarios generation of IKC spreading for forecasting issues.

ET proved to be a good predictor of IKC Pr, with no significant conditioning by altitude. EOS seems to operate good predictions of Pr at lower altitudes, while LOS at medium-high ones. This probably depends on wildlife dynamics, that in autumn and at the beginning of winter, looks for grassland (food) especially at lower altitudes where, in that period, grass is not covered by snow and is wetter. In the case of SOS, a possible interpretation can rely on the fact that, if vegetative season lasting increase, animals descent from mountain to valley can be postponed, thus inducing a higher probability of interaction between potential guests and sick animals at higher altitudes. This certainly can increase also the probability of exposure of animals to disease.

With reference to ET it is mainly related to Pr at lower altitudes where micro-local temperature and humidity favor vegetation growth and limit soil drought, determining a higher attractiveness for chamois. At higher altitudes ET capability of predicting Pr becomes weaker probably due to a lower availability of biomass and a shorter phenological season.

Authors are conscious that this work just introduces a new way to manage wildlife health problem and cannot be retained conclusive. In the nearer future, more disaggregated investigations should be done, and other areas possibly considered. Nevertheless, the proposed approach is sufficiently innovative in the context of wildlife veterinary and, we hope, could open a new interesting trend to map wildlife diseases and related zoonosis risk associated with the interaction between wild animals and domestic ones. A radical change is expected also by technicians and institutional subjects in their ordinary procedures for recording and manage ground data. In fact, the greatest limit to expand and more focus this research relied on the format of ground data that could be obtained only aggregated at regional level with no information concerning the specific place where each analyzed animal was found. We invite all involved players to carefully consider the possibility of georeference every ground observation that comes to their laboratories. Georeferencing of ground data is at the basis of an effective and reliable spatial based approach like the one here proposed, where EO data (especially if available over a long time span) play a crucial role. Anyway, this work proved that spatially based forecasting models can be reasonably calibrated for generating maps of risk concerning wildlife diseases and zoonosis spreading in a certain area. Moreover, it showed that relationship between IKC and PMs/ET are probably chancing in terms of strength; in fact, we demonstrated that all the considered predictors are suffering from a significant change possibly related to the ongoing climate change. Consequently, we expect that future approaches should more properly rely on contemporary data spatially distributed in place of aggregated data temporally distributed like the one we processed for this work.” (please see lines 607-691)

 

Regarding to the model calibration, the IKC prevalence data is from 2009 – 2019. During the calibration process it has been followed the procedure explained in Analysis 2 section and adopted eq. 3. It has been found a single one outlier only in some cases (FAVs). Anyway, if existed, it was reported into each graph with a red orange point and in the caption section below.

(please see lines 555-563 Analysis 2 section into the Results).

 

Regarding to the 2nd order polynomial regression the referee is absolutely right. IKC Pr was tested against all the computed metrics by 2nd order polynomial regression and p-value were calculated. This analysis has permitted to detect which PMs and ET in which altimetry bands area are really more sensitive to IKC Pr. We have made a mistake in the methods section reporting Pearson’s that instead it was used in analysis 1, we really apologize for it and made the correction. (please see lines 446-460 Analysis 2 section).

 

 

The Yearly Cumulative ET seems to be a good predictor and respond better than average ET, for this reason has been considered. Anyway, as wisely indicated by the referee it wasn’t mentioned in the methods section. For this reason, following the referee suggestion, Yearly Cumulative ET has been included as a variable, in the methods, in particular in the Analysis 2 section as follow: “After demonstrating that significant trends could be recognized affecting PMs and ET values from MOD13Q and MOD16A2 datasets, respectively, we tested their potential correlation with IKC Pr at regional level. As far as IKC/ET comparison was concerned the Yearly Cumulative ET was calculated. The a priori hypothesis was that a change in environmental conditions could drive to a change in IKC occurrences (Pr value). It is worth to remind that Pr data were supplied aggregated at regional level, while PMs and ET measures were mapped at pixel level over the whole area. Consequently, all deductions refer to general trends that could be possibly improved if more distributed and geolocated data of IKC Pr were available.

IKC Pr was tested against all the computed metrics and modelled by a 2nd order polynomial regression.” (please see lines 446-460 Analysis 2 section).

 

 

The sentence in which we said that the study has a small number data points to be able to provide robust forecasting, has been revised as follow “Authors are conscious that this work just introduces a new way to manage wildlife health problem and cannot be retained conclusive. In the nearer future, more disaggregated investigations should be done, and other areas possibly considered. Nevertheless, the proposed approach is sufficiently innovative in the context of wildlife veterinary and, we hope, could open a new interesting trend to map wildlife diseases and related zoonosis risk associated with the interaction between wild animals and domestic ones. A radical change is expected also by technicians and institutional subjects in their ordinary procedures for recording and manage ground data. In fact, the greatest limit to expand and more focus this research relied on the format of ground data that could be obtained only aggregated at regional level with no information concerning the specific place where each analyzed animal was found. We invite all involved players to carefully consider the possibility of georeference every ground observation that comes to their laboratories. Georeferencing of ground data is at the basis of an effective and reliable spatial based approach like the one here proposed, where EO data (especially if available over a long time span) play a crucial role. Anyway, this work proved that spatially based forecasting models can be reasonably calibrated for generating maps of risk concerning wildlife diseases and zoonosis spreading in a certain area. Moreover, it showed that relationship between IKC and PMs/ET are probably chancing in terms of strength; in fact, we demonstrated that all the considered predictors are suffering from a significant change possibly related to the ongoing climate change. Consequently, we expect that future approaches should more properly rely on contemporary data spatially distributed in place of aggregated data temporally distributed like the one we processed for this work.” (please see lines 670-691).

 

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

This manuscript contains some interesting data and the presentation has been considerably improved. Table 4 needs a little more explanation especially with regard to stated prevalence vs number of +ve to IKC. I suggest that the introduction could still be shortened with some of the text better situated in the discussion. There is enough data in the paper to focus on IKC vs the range of other disease issues mentioned. Many of the grammatical errors have been addressed but the final version still needs proof reading because the text remains misleading in places. The latter continues to impact the potential quality of the publication.

Author Response

Response to Reviewer 2 Comments

 

We would like to thank reviewers for their appropriate comments and helpful suggestions that have been carefully considered. Majority of provided suggestions highlighted gaps in the text and were really useful to improve, we hope, paper quality.

In particular, the synthesis of reviewers’ comments suggested a deep revision in paper organization and harmonization. Consequently, you will find some structural changes aimed at simplifying paper reading and make content more effective.

All comments were carefully evaluated and for the most of them corrections and integrations have been provided.

 

 

Point 1: This manuscript contains some interesting data and the presentation has been considerably improved. Table 4 needs a little more explanation especially with regard to stated prevalence vs number of +ve to IKC. I suggest that the introduction could still be shortened with some of the text better situated in the discussion. There is enough data in the paper to focus on IKC vs the range of other disease issues mentioned. Many of the grammatical errors have been addressed but the final version still needs proof reading because the text remains misleading in places. The latter continues to impact the potential quality of the publication.

Response 1: Firstly, we would like to thank the reviewer for his/her appropriate previous and present suggestions.

Concerning table 4, as wisely suggested by the referee, a deeper explanation was added in the caption as it follow: “Faunistic season is intended from September (previous year) to August (next year); consequently IKC disease prevalence values were re-assigned at year level assuming the period January - December as reference time to perform analysis 2. It’s worth to remind that positive IKC column refers to chamois that have been detected positive to the IKC considering the overall number of samples analyzed per each year reported in the fourth column after performing a PCR analysis (please see Veterinary Ground Samples Section).” (Please see lines 385-393)

 

 

The introduction was further shortened, as suggested by the referee, by moving some parts to the discussions as it follows.

 

[Introduction] “Geomatics and satellite remote sensing represent a useful analysis tool in several technical-scientific fields [1]. Nowadays remote sensing is widely used in many fields like agronomy, forestry and environment, in general. Nevertheless, veterinary and faunistic-related applications are still limited and, often, characterized by an improper use of Earth Observation (EO) data, that makes desirable a significant improvement [2].

On the international scene, only few research groups are currently properly exploiting all the potentialities that Geomatics (remote sensing included) and digital geographical data could offer to the veterinary sector. The most of works concern parasitology and virology for etiological/epidemiological studies [3],diagnosis and medical history [4]. In these context, satellite remote sensing has assumed a great interest in the last years [5]; EO data are proficiently used to feed meteorological and climatological models with the aim of generating predictive scenarios of zoonosis spread and outbreaks [6].

Initially, the most of studies in the veterinary and health sectors used remote sensing to describe environmental conditions; this occurred especially with reference to malaria in Africa and Asia [7; 8]. Presently, epidemiologists are adopting remote sensing to investigate a variety of vector-borne diseases. Associations between remote sensing-derived environmental variables (e.g. temperature, humidity, land cover, etc.) and vector density are used to map and characterize vector habitats [9]. The basic idea is that remotely sensed data can contain dynamic predictors of Earth’s processes suitable for describing niche preferences of some medically important host diseases mechanisms. Moreover, because of their continuity of acquisition, remotely sensed data provide a synoptic representation of environment at proper spatial and temporal scales [10].

Meteorological and EO data are often jointly used for diseases analysis. For example, outbreaks of diarrheal disease and, specifically, cholera were analyzed by a new modeling approach based on satellite data to produce cholera risk maps in several regions of globe [8], supporting the idea that the ongoing EO technology transfer is making possible to investigate new patterns in a systemic point of view [11].

Given the veterinary and public health impact of vector-borne diseases, there is a clear and immediate need to map and monitor local landscape attitude to encourage emergence and spread of these diseases. Current approaches for predicting disease risks are mostly neglecting key features of landscape related to the functional habitat of vectors (or hosts) and, hence, of the pathogen [12].

A global satellite-based monitoring of proper climate variables could help to map occurring anomalies with the aim of predicting spatial distribution of risk related to emergence and propagation of disease vectors. Such information could provide sufficient lead-time for outbreak prevention and potentially reduce burden and spread of ecologically coupled diseases.

Additionally, remote sensing could have an important role in the comprehension of patho-system dynamic. With reference to the so called “disease triangle”, including host, pathogens/vectors and environment, remote sensing, and geomatics in general, could support scientists and decision-makers to better understand the role of environmental patterns and, therefore, explore its complex relations with the other parts of patho-system [13; 14].

A good example is represented by atmospheric pollution that was recognized to increase sensitivity to pulmonary diseases, as the last pandemic event, coronavirus (SARS-CoV-2), has suggested [15].

GIS studies about endo- and ecto-parasitoses of veterinary interest, with particular reference to zoonoses agents, represent today the greatest contribution to veterinary and faunistic sectors [16]. For many years the World Organization for Animal Health located in Paris (OIE) and the World Health Organization (WHO) in Geneva have been underlining the importance of Geomatics and remote sensing applications [17] in the "One Health" perspective.

This work, with reference to a regional case study, investigates remote sensing potentialities for describing relationships between environment and diseases affecting wildlife at landscape level. Moreover, it is intended to describe the effects of climate change onto the vegetation component, with special concern about pastures. The study area corresponds to the entire Aosta Valley Region located in the Italian Western Alps. In particular, a new analysis approach is presented to operate at landscape level to analyze if and how environmental factors could condition the occurrence of infectious keratoconjunctivitis (IKC, Mycoplasma conjunctivae) in chamois. IKC is a contagious disease for domestic and wild ruminants (Caprinae and Ovinae) [18]. In chamois, the disease can be serious [19] and, as in other wild ruminants, blindness can occur [20], with consequent death of the animal from trauma (e.g. fall from cliffs or starvation) [21]. The period of mountain pasture is risky for the potential contact between domestic and wild infected animals; along the years, several outbreaks have been reported in wild ungulates in the Alps [22] and this is the reason that makes monitoring / surveillance plans still active. IKC caused by Mycoplasma conjunctivae is a complex disease of domestic and wild Caprinae, with great variations in the clinic-pathological and epidemiological picture. In wildlife, IKC is sometimes associated with high mortality [23; 24]. It has been suggested that the pathogenesis of IKC is influenced by host predispositions, virulence of M. conjunctivae strains, secondary infections, and environmental factors [25]. Sex and age imbalance in affected populations were observed in severe outbreaks [26], indicating that age and social behavior, including sexual segregation, may be important risk factors. Differently, differences in virulence between different strains do not seem to play a major role; mycoplasmal load is obviously associated to the presence and severity of signs. However, the driver of mycoplasmal multiplication in the host is unknown. Environmental factors might have a role, regarding both the expression of the disease in individual cases and the onset of an outbreak in a population [24]. The underlying hypothesis of this work is that remote sensing could support comprehension of the role of environmental patterns in conditioning IKC patho-system, and related pathologies, as for other diseases. Altitude, air quality, and UV light have been discussed as possible predisposing factors for IKC in wild ungulates along with overcrowding [27]. Multiple outbreaks of IKC in Alpine ibex and Alpine chamois populations have been described in literature [28]. Different outbreaks of infectious keratoconjunctivitis (IKC) affecting alpine chamois and ibex in the western and central Swiss Alps and Aosta Valley were recorded in the period 2001-2019 [29]. Between the years 2001-2003, in Switzerland, Mycoplasma conjunctivae was identified from conjunctival swabs by means of a nested PCR in 27 of the 28 chamois tested. The outbreaks occurred in an area covering 1590 km². Deep valleys acted as a barrier to the spread of the disease. Many of the affected animals were juveniles, and more females than males died of IKC. The disease was more common during the summer and autumn.  In some outbreaks, mortality can reach 30 per cent, as, for example, in chamois in Italy, France and Switzerland, and hundreds of chamois may die. Major outbreaks were recorded in 2001-2003 [30] and 2016-2018 [31]. With these premises, in this work, two types of analysis were performed: one aimed at exploring, by remotely sensed data, phenological metrics (PMs) and evapotraspiration (ET) trends of vegetation; one investigating correlation between PMs and ET versus IKC prevalence. PMs/ET analysis was based on TERRA MODIS image time series ranging from 2000 to 2019. Ground data about IKC were available for a shorter time range: 2009 - 2019. Consequently, PMs and ET trends investigation were done for the whole times range (2000-2019); conversely, correlation analysis was achieved with reference to the 2009-2019 period.” (Please see lines 45-152)

 

[Discussions] “The functional roles of domestic and wild host populations in infectious keratoconjunctivitis (IKC) epidemiology have been extensively discussed claiming a domestic reservoir for the more susceptible wild hosts; in the most of cases all deductions were based on limited data.

With the aim to better assess IKC epidemiology in complex host-pathogen alpine systems, the long-term infectious dynamics and molecular epidemiology of Mycoplasma conjunctivae has been investigated in all host populations from different areas in the Pyrenees and Occidental Alps.

Between the years 2000-2019, it was consistently detected in Pyrenean and Alpine chamois (Rupicapra p. pyrenaica) populations, as well as in sheep flocks, and occasionally in mouflon (Ovis aries musimon) from the Pyrenees; statistically associated with ocular clinical signs only in chamois. Chamois populations showed different infection dynamics with low but steady prevalence (4.9%) and significant yearly fluctuations (0.0%– 40.0%) between the period 2008-2015 [27, 28]. Persistence of specific M. conjunctivae strain clusters in wild host populations is demonstrated for six and nine years. Cross-species transmission between chamois and sheep and chamois and mouflon were also sporadically evidenced. In Switzeland, the chamois affected by IKC were found at altitudes between 550 and 3200 m. The estimated overall mortality was less than 5 per cent, but more than 20 per cent have probably died locally [29]. Host population characteristics and M. conjunctivae strains resulted in different epidemiological scenarios in chamois, ranging from the fading out of the mycoplasma to the epidemic and endemic long-term persistence. These findings highlight the capacity of M. conjunctivae to establish diverse interactions and persist in host populations, also with different transmission conditions. Overall, independent M. conjunctivae sylvatic and domestic cycles occurred at the wildlife-livestock interface in the alpine ecosystems with sheep and chamois as the key host species for each cycle, and mouflon as a spill-over host. Although outbreaks of IKC have been described in Austria, France, Italy, Slovenia and Switzerland, descriptive studies of the role of environmental patterns and to model the outbreaks on a large scale have often been incomplete, owing to the difficulty of detect fundamentals pattern that can affecting IKC spread in chamois that living in remote, inaccessible mountain regions. Under this scenario the remote sensing techniques and EO data can give certainly a huge hand in the understanding and development of possible forecasting models, as we have tried to do in the present work. With these premises the present study, was intended to explore and propose a method based on free accessible EO data to partially close the above-mentioned knowledge gap. 

In Aosta Valley (NW Italy) PMs and ET (as measured from the above mentioned EO data) proved to significantly changed their values in the last 20 years, with a continuous progressive trend observable for all of them. In terms of strength of changes, an average delay of EOS was observed of about 2.6 days, independently from the altitude class. SOS proved to averagely anticipate of about 2 and 3 days per year at lower (< 2000 m) and higher (> 2000 m) altitudes, respectively. Consequently, LOS is enlarging of about 4.7 and 6.5 days per year at lower (< 2000 m) and higher (> 2000 m) altitudes, respectively. While looking at the entire period (2000-2019) MAXVI proved to be significantly changing, showing a positive variation (about +0.09) at lower altitude and no variations at higher one. This can be explained admitting that at lower altitudes, in Aosta Valley, grasslands and pastures are often irrigated. Consequently, farmers can vary water release regimes to face climate change effects (higher temperatures, in particular) with the result of moving forage yields (that NDVI is a predictor of) to higher values.

Differently, where more natural (not managed) systems are located (higher altitudes) the increasing of yearly MAXVI can be only related to glacier melting that could compensate the increasing of water requirement (as confirmed by the ET analysis) by vegetation: glaciers are, in fact, dramatically reducing in Aosta Valley. Moreover, another compensating action could come from the surrounding forest areas that have been proved to tolerate summer heatwaves.

With reference to ET, a significant increasing trend was observed, independently from altitude.  Eight days water requirement from vegetation appears to averagely increase of about 0.05 Kg·m^-2 (about 0.5%) every year for a total increase of about 1 Kg·m^-2 in 20 years (2000-2019), corresponding to a percentage difference in water requirement from vegetation of about 8%. This could be possibly explained by the increasing of biomass production (well represented by MAXVI) and by the enlargement of the growing season, that, consequently make vegetation needing more water yearly.

As far as PMs/ET and IKC Pr correlation is concerned, some interesting findings came out.

A 2^nd order polynomial model was found to well approximate the most of relationships, making possible to support scenarios generation of IKC spreading for forecasting issues.

ET proved to be a good predictor of IKC Pr, with no significant conditioning by altitude. EOS seems to operate good predictions of Pr at lower altitudes, while LOS at medium-high ones. This probably depends on wildlife dynamics, that in autumn and at the beginning of winter, looks for grassland (food) especially at lower altitudes where, in that period, grass is not covered by snow and is wetter. In the case of SOS, a possible interpretation can rely on the fact that, if vegetative season lasting increase, animals descent from mountain to valley can be postponed, thus inducing a higher probability of interaction between potential guests and sick animals at higher altitudes. This certainly can increase also the probability of exposure of animals to disease.

With reference to ET it is mainly related to Pr at lower altitudes where micro-local temperature and humidity favor vegetation growth and limit soil drought, determining a higher attractiveness for chamois. At higher altitudes ET capability of predicting Pr becomes weaker probably due to a lower availability of biomass and a shorter phenological season.

Authors are conscious that this work just introduces a new way to manage wildlife health problem and cannot be retained conclusive. In the nearer future, more disaggregated investigations should be done, and other areas possibly considered. Nevertheless, the proposed approach is sufficiently innovative in the context of wildlife veterinary and, we hope, could open a new interesting trend to map wildlife diseases and related zoonosis risk associated with the interaction between wild animals and domestic ones. A radical change is expected also by technicians and institutional subjects in their ordinary procedures for recording and manage ground data. In fact, the greatest limit to expand and more focus this research relied on the format of ground data that could be obtained only aggregated at regional level with no information concerning the specific place where each analyzed animal was found. We invite all involved players to carefully consider the possibility of georeferencing every ground observation that comes to their laboratories. Georeferencing of ground data is at the basis of an effective and reliable spatial based approach like the one here proposed, where EO data (especially if available over a long-time span) play a crucial role. Anyway, this work proved that spatially based forecasting models can be reasonably calibrated for generating maps of risk concerning wildlife diseases and zoonosis spreading in a certain area. Moreover, it showed that relationship between IKC and PMs/ET are probably chancing in terms of strength; in fact, we demonstrated that all the considered predictors are suffering from a significant change possibly related to the ongoing climate change. Consequently, we expect that future approaches should more properly rely on contemporary data spatially distributed in place of aggregated data temporally distributed like the one we processed for this work.” (Please see lines 495-583)

 

Least but not last, a further English comprehensive editing has been done taking into account the suggestion proposed by the referee. In order to avoid to report here all the changes performed, we kindly suggest the reviewer to see the revised paper.

 

Author Response File: Author Response.docx

Reviewer 3 Report

The authors have updated the manuscript to reflect the changes I suggested. The clarity of the manuscript is improved and the methods used for Analysis 2 are clear and appropriate. 

Author Response

Response to Reviewer 3 Comments

 

We would like to thank reviewers for their appropriate comments and helpful suggestions that have been carefully considered. Majority of provided suggestions highlighted gaps in the text and were really useful to improve, we hope, paper quality.

In particular, the synthesis of reviewers’ comments suggested a deep revision in paper organization and harmonization. Consequently, you will find some structural changes aimed at simplifying paper reading and make content more effective.

All comments were carefully evaluated and for the most of them corrections and integrations have been provided.

 

Point 1: The authors have updated the manuscript to reflect the changes I suggested. The clarity of the manuscript is improved and the methods used for Analysis 2 are clear and appropriate.

 

Response 1: We would like to thank the reviewer for his/her comment. A further English comprehensive editing has been done taking into account also the suggestion proposed by others referee. In order to avoid to report here all the changes performed, we kindly suggest the reviewer to see the revised paper.

 

 

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

This paper examines how geomatics and satellite remote sensing might offer a useful tool in predicting outbreaks of infectious keratoconjunctivitis (IKC) in chamois (Rupicapra rupicapra L.) in Aosta Valley (NW Italy).

Although some of the results are interesting the paper is currently unfocussed and very hard to follow. The introduction would be improved if it focused on IKC and the factors that might impact the occurrence of this clinical disease in Chamois in this region of Italy. I think that it is already well accepted that geomatics and remote sensing can be useful in modeling the probability of disease outbreaks, especially for outbreaks of vector borne diseases such as malaria which have significant links to changes in climate. Such models, however, need to take into account  the disease ecology and host factors specific for a particular disease outbreak scenario. Although this is mentioned it isn't clear from the methods section how this information was incorporated in the interpretation of the results. Quite a bit of the results section seems to include comments that might be better placed in the discussion.

The paper currently requires comprehensive editing.

Reviewer 2 Report

The manuscript “Geomatics and EO data to support wildlife diseases assessment at landscape level: a pilot experience to map infectious keratoconjunctivitis in chamois in Aosta Valley (NW Italy)” reports two analysis performed by the authors. The first, consider the favourable CORINE classes for chamois presence and analyse their evolution in the last two decades in terms of some phenological metrics based on NDVI and evapotranspiration. In the second analysis, some of these metrics were associated to disease regional cases (prevalence).

Although well written, the manuscript present a lot of redundancies in description of some parts; moreover all relevant parts of the study are described along the manuscript, but they are dispersed across the sections. A reorganisation of the content is needed. This reviewer would avoid to merge together results and discussion, and would report in the conclusions only the final statement of the manuscript.

The paper is recommended for publication on Remote Sensing, however some major revisions and clarifications are needed, as follows.

Overall comments

• Although well described, the manuscript present a lot of redundancies in descriptions of the EO products, in the formulas, in the introduction with examples out of the scope of the paper. The reader (and myself) get lost in these details, instead of focusing on the novelty and relevant information of the study and of the disease under investigation.
• The aim of the paper with the two analysis should be clearly reported in the abstract and at the end of the introduction section.
• The description of EO data used, needs to be clearer in terms of what is already available (avoiding long descriptions that the reader can access in the dedicated webpages) and what are the elaborations that the authors perform on them. In general, first describe the product with its relevant information (source, resolutions), then describe the manipulation you performed.
• Considering the high number of acronyms present in the manuscript, please consider to add an abbreviation list, to facilitate the reader.
• The disease cases refers to the faunistic season that is from September to August of next year, while the phenological metrics refer to yearly data (from January to December). How did you integrate the two different periods? It results unclear to this reviewer. Moreover, the disease cases are available at regional level, and the phenological metrics are based on altitude: how do you integrated these two pieces of information?

See here some detailed suggestions:

 

Introduction. Great part of the introduction results out of scope to this reviewer. I can understand you want to put relevance to remote sensing and geomatic application, and I agree on this, but so many details on Africa context and different disease (either because of the virus or the hosts) seem not relevant here and confuse the reader. Consider to remove or sum up the lines 51-64, 70-86, 93-104. These parts can be substituted with some examples of application in the same (or similar) disease or to a disease affecting the same animal species or involving the same environmental characteristics, almost in a similar region (not Africa and Malaria). Useful references can be:

Kalluri S, Gilruth P, Rogers D, Szczur M (2007) Surveillance of arthropod vector-borne infectious diseases using remote sensing techniques: A review. PLoS Pathog 3(10): e116. doi:10.1371/journal.ppat.0030116

McIntyre, K.M., Setzkorn, C., Hepworth, P.J., Morand, S., Morse, A.P., Baylis, M., 2017. Systematic Assessment of the Climate Sensitivity of Important Human and Domestic Animals Pathogens in Europe. Scientific Reports 7, 7134. https://doi.org/10.1038/s41598-017-06948-9

Jamison, A., Tuttle, E., Jensen, R., Bierly, G., Gonser, R., 2015. Spatial ecology, landscapes, and the geography of vector-borne disease: A multi-disciplinary review. Applied Geography 63, 418–426. https://doi.org/10.1016/j.apgeog.2015.08.001

And consider to include these interesting lectures and applications of remote sensing in veterinary field:

http://dx.doi.org/10.1016/j.apgeog.2015.08.001

DOI:10.1371/journal.pone.0140915

https://doi.org/10.1073/pnas.1521657113

https://doi.org/10.1371/journal.pone.0146024

https://doi.org/10.1186/1756-3305-7-302

doi:10.1371/journal.pone.0112491

https://doi.org/10.1111/brv.12149

https://doi.org/10.1186/1476-072X-13-26

https://doi.org/10.4081/gh.2013.72

The introduction to the disease itself is almost completely missing. Please add sentences describing the disease and relate it to the geographical context or surrounding areas. This reviewer suggests to expand the lines 121-132, that at the moment can be considered the introduction section. Explain the concern about the disease, the size and characteristics of the animal population affected, why the environment is relevant and which studies already exist associating environment and keraconjuntivitis, where they are located, their insights and main limitations. And then suggest to approach the problem with landscape measures, explaining the rationale behind this approach and providing some examples of its application elsewhere. Please evaluate the following references and consider to add them to the manuscript for completeness:

1. Dynamics of an Infectious Keratoconjunctivitis Outbreak by Mycoplasma conjunctivae on Pyrenean Chamois Rupicapra p. pyrenaica. https://doi.org/10.1371/journal.pone.0061887
2. Long-term dynamics of Mycoplasma conjunctivae at the wildlife-livestock interface in the Pyrenees

https://doi.org/10.1371/journal.pone.0186069

3. Tschopp R, Frey J, Zimmermann L, Giacometti M. Outbreaks of infectious keratoconjunctivitis in alpine chamois and ibex in Switzerland between 2001 and 2003. doi:10.1136/vr.157.1.13
4. Is the development of infectious keratoconjunctivitis in Alpine ibex and Alpine chamois influenced by topographic features? https://doi.org/10.1007/s10344-012-0651-1

Materials and methods

Lines 139-140 and figure 2. Please, consider to report on the map, the boundaries of the three protected areas you mention, if relevant for the paper.

Lines 147-150 seem a repetition of lines 140-142. Please consider to remove them.

Line 155. I guess the figure here is the number 3 (not 4).

Lines 165-166. The sentence is incomplete.

Line 183. A colon is probably missing after “was used”.

Lines 201-204 and table 2. Please explain (not only in the caption of the table) why the selected CORINE classes, namely 231, 321, 322, are the only ones considered the favourable to host animals. Why not the 324, or the 333? Answer to this question is reported at lines 294-304, in methodology section; I suggest to move here the description and selection of the CORINE classes.

Table 2. Please consider to add a column with the percentage of surface that each class occupy in Aosta Valley Region.

I suggest to move part of the description of Aosta Valley region in a supplementary doc, or reduce this content to what is relevant for the analyses. Consider to add a map with the CORINE classes relevant for the study.

Veterinary ground samples

Lines 209-219. Please report here that the geographical coordinates of the places where animals are collect is not available, and that the study was performed at regional level.

Line 210. Please explain the meaning of the acronym RFD.

Lines 239-240. All samples collected in the whole region (Aosta) and in the whole period (2009-2019) were used in the analyses? What the “RFD subsample” refers to? Please clarify the sentence, as it is not clear to this reviewer.

EO and Geographical Digital Data

In this paragraph, it is unclear if the three products, i.e. NDVI, Evapotranspiration and altitude, were resampled at the same spatial and temporal resolutions, considering their native different ones (30, 250, 500 meters, 16 days an 8 days for temporal resolutions).

Line 255. A full stop is probably missing after “per pixel basis”.

Lines 247-264 Please use the same name for all the occurrences of the text “MOD13Q1-v.6” product throughout the paragraph. And please use the same symbols in the formula at lines 261 and lines 263-264.

Lines 247-264. Please reorganise the sentences as it is not clear to this reviewer if the product downloaded from Google Earth Engine are ready to use, with the filter Savitzky-Golay already applied, or if the authors started with spectral bands (NIR and RED), then applied the filter, followed by the formula (equation 2), to obtain the NDVI rasters. In both cases, please specify the software used to manipulate the rasters (there is a reference to ENVI software, but I am not sure it was used for the analyses).

Line 266. Please check the dates of evapotranspiration product: from the website https://developers.google.com/earth-engine/datasets/catalog/MODIS_006_MOD16A2 it seems that the product availability starts at 01.01.2001.

Lines 268-272. The text reported is the same of the description reported in the page https://developers.google.com/earth-engine/datasets/catalog/MODIS_006_MOD16A2: please mention the page or modify the sentence.

Line 271. Please clarify how the references 35 and 36 are used to calculate the evapotranspiration.

Methodology

Lines 283-286. Please consider to better explain the objective of analysis 1, as to this reviewer it is not clear. As far as I understood, it is an environmental – climatic analysis, independent on the disease, based on the years 2000-2019. What is the meaning of phenological metrics? Add here a list of the metrics, reported at lines 319-325. 

Line 283. Please add a coma after the word “factors”.

Lines 294-306. Please integrate this part in Materials and Methods, at the end of the study area section.

Line 308. Please report only the software and the method that were used in the analyses, i.e. QGIS or SAGA GIS?  Why both?

Line 309. Please consider to add a map, showing the FAV1, FAV2, FAV3 areas.

Line 326. Which version of the TIMESAT software was used in this manuscript?

Lines 326-333. Please consider to rephrase the sentence, simply explaining what the software does. For example: “the software iteratively fits smooth mathematical functions to yearly time-series of noisy satellite data at per-pixel level, to derive the best smoothed approximation of the NDVI along the year. Then, key phenological metrics (beginning and end of the growing season, length of the season, amplitude, integrated value, asymmetry of the season etc.) are extracted for each pixel.”

Line 345. Something is missing to MOD13Q. And add “s” to dataset.

Results

Lines 361-363. In the table, please add the number of the samples (the 2% prevalence can derive from 2/100 or 200/10000 and the reliability of the value is different).

Lines 366-376 this part is material and methods.

Figure 5. Please report the acronyms in the caption of the figure (SOS, DOY, etc.). And report the legend of the color of the points (red are the outliers?).

Line 391. Figure 8 is not existing in the manuscript, probably the reference is to figure 5.

Table 5. What is the meaning of SOS (DOY) when it is equal to -38.76. Is it the Day Of the Year (going from 0 to 365)? Please explain or modify.

Table 6. If the disease cases are available at regional level, as you say, how do you divided them by altitude, in FAV1, FAV2, FAV3? In the table, please highlight the numbers that are statistically significant.