Skip to main content
SearchLoginLogin or Signup

Review 1: "Ecological Impacts of Climate Change will Transform Public Health Priorities for Zoonotic and Vector-borne Disease"

The authors agree that this preprint is potentially informative, but that much more extensive models should be utilized due to the extreme complexity of the question and range of diseases studied.

Published onMay 20, 2024
Review 1: "Ecological Impacts of Climate Change will Transform Public Health Priorities for Zoonotic and Vector-borne Disease"
1 of 2
key-enterThis Pub is a Review of
Ecological impacts of climate change will transform public health priorities for zoonotic and vector-borne disease
Ecological impacts of climate change will transform public health priorities for zoonotic and vector-borne disease
Description

Abstract Climate change impacts on zoonotic/vector-borne diseases pose significant threats to humanity1 but these links are, in general, poorly understood2. Here, we project present and future geographical risk patterns for 141 infectious agents to understand likely climate change impacts, by integrating ecological models of infection hazard (climate-driven host/vector distributions and dispersal3,4) with exposure (human populations) and vulnerability (poverty prevalence). Projections until 2050, under a medium climate change (Representative Concentration Pathway (RCP) 4.5), show a 9.6% mean increase in endemic area size for zoonotic/vector-borne diseases globally (n=101), with expansions common across continents and priority pathogen groups. Range shifts of host and vector animal species appear to drive higher disease risk for many areas near the poles by 2050 and beyond. Projections using lower climate change scenarios (RCP 2.6 & 4.5) indicated similar or slightly worse future population exposure trends than higher scenarios (RCP 6.0 & 8.5), possibly due to host and vector species being unable to track faster climatic changes. Socioeconomic development trajectories, Shared Socioeconomic Pathways (SSPs), mediate future risk through a combination of climate and demographic change, which will disrupt current, regional patterns of disease burden. Overall, our study suggests that climate change will likely exacerbate global animal-borne disease risk, emphasising the need to consider climate change as a health threat.One Sentence Summary Climate change and socio-economic development dictate future geographical areas at risk of zoonotic and vector-borne diseases.

RR:C19 Evidence Scale rating by reviewer:

  • Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.

***************************************

Review: The idea is interesting and merits publications once some methodological flaws are addressed. The species distribution modeling has risk of exaggerated predictions under future climate. The model ensemble and risk estimates are innovative.

The present study proposed maps of disease transmission risk for a series of infectious diseases, of animal origin, affecting humans. The contribution is timely and adds to the growing body of literature on climate change and infectious disease forecasts. Species distribution modeling is the core analytical framework of the article. Results from the species distribution modeling are then used for subsequent analyses. Because severe flaws were noted in the species distribution modeling, there are concerns regarding the robustness of the results, the accuracy of the inferences and conclusions, and the risk of misleading recommendations. Revising the methods employed in the species distribution modeling could generate different results. As such, I recommend conducting more exhaustive models using well accepted modeling protocols to reduce, detected, or manage uncertainty. First, model selection should be comprehensive by including diverse study area extents of background point densities for a more robust contrast between observed and available environments. Second, model evaluation should include robust metrics of model fit and model prediction. Finally, projection of models to different periods should consider risk of model extrapolation and presence of novel climates. Below some observations. In the current form, I have concerns regarding the ecological signals captured for reservoirs and vectors to infer future transmission risk.

Observations:

Host-vector data collection has many decisions and assumptions not supported by references.

Maxent is not a presence-absence model, it is a presence background model. The interpretation of the algorithm as presence-absence, may be linked to the low number of background points used (e.g., from 100 to 5000), which influences the model prediction. Maxent contrasts density of environmental conditions from the occurrences and the study area. A low number of points may underrepresent the conditions in the study area and will result in inaccurate estimation of suitable pixels. 

It is unclear why elevation was included considering that elevation is not expected to change under climate change, and information of elevation may be contained in other variables, e.g., temperature.

Cross validation:

10% may be a too low sample size for a robust model evaluation. I assume the model evaluation data was generated using random sampling. This does not ensure statistical independence between data for model calibration and evaluation. Finally, there is strong evidence demonstrating that AUC is a weak metric to differentiate bad from good models. Instead, other metrics based on model fit and significance could be used for model selection. 

Elevation was used in some models and altitude was used in other models. This needs to be revised since both variables have different metrics and values.

Model projections to future climates using Maxent were unclear. It seems that extrapolation and clamping were used. This decision tends to generate the distribution predicted by allowing the models to predict into novel environments without control on the maximum conditions tolerated in any variable.

References:

  1. Fourcade Y, Besnard AG, Secondi J. Paintings predict the distribution of species, or the challenge of selecting environmental predictors and evaluation statistics. Glob Ecol Biogeogr (2017) 27:245–256. doi:10.1111/geb.12684

  2. Golicher D, Ford A, Cayuela L, Newton A. Pseudo-absences, pseudo-models and pseudo-niches: Pitfalls of model selection based on the area under the curve. Int J Geogr Inf Sci (2012) 26:2049–2063. doi:10.1080/13658816.2012.719626

  3. Cobos ME, Peterson AT, Barve N, Osorio-Olvera L. kuenm: A dynamic R package for detailed development of ecological niche models using Maxent. PeerJ (2019) 7:e6281. doi:10.7717/peerj.6281

  4. Anderson RP. A framework for using niche models to estimate impacts of climate change on species distributions. Ann N Y Acad Sci (2013) 1297:8–28. doi:10.1111/nyas.12264

  5. Araújo MB, Anderson RP, Barbosa AM, Beale CM, Dormann CF, Early R, Garcia RA, Guisan A, Maiorano L, Naimi B, et al. Standards for distribution models in biodiversity assessments. Sci Adv (2019) 5:1–12. doi:10.1126/sciadv.aat4858

  6. Mesgaran MB, Cousens RD, Webber BL. Here be dragons: A tool for quantifying novelty due to covariate range and correlation change when projecting species distribution models. Divers Distrib (2014) 20:1147–115

Comments
0
comment
No comments here
Why not start the discussion?