On the inadequacy of species distribution models for modelling the spread of SARS-CoV-2: response to Araújo and Naimi

The ongoing pandemic of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is causing significant damage to public health and economic livelihoods, and is putting significant strains on healthcare services globally. This unfolding emergency has prompted the preparation and dissemination of the article “Spread of SARS-CoV-2 Coronavirus likely to be constrained by climate” by Araújo and Naimi (2020). The authors present the results of an ensemble forecast made from a suite of species distribution models (SDMs), where they attempt to predict the suitability of the climate for the spread of SARS-CoV-2 over the coming months. They argue that climate is likely to be a primary regulator for the spread of the infection and that people in warm-temperate and cold climates are more vulnerable than those in tropical and arid climates. A central finding of their study is that the possibility of a synchronous global pandemic of SARS-CoV-2 is unlikely. Whilst we understand that the motivations behind producing such work are grounded in trying to be helpful, we demonstrate here that there are clear conceptual and methodological deficiencies with their study that render their results and conclusions invalid.What follows is a response to the Araújo and Naimi article centered around three main criticisms:1) Given the fact that SARS-CoV-2 has a primary infection pathway of direct contact, it is in an active spreading phase, and remains largely underreported in the Global South, it represents an inappropriate system for analysis using the SDM framework.2) Even if we were to accept that an SDM framework would be applicable here, the methodology presented in the article strays far from best-practice guidelines for the application of SDMs.3) The dissemination strategy of the authors failed to respect the frameworks of risks adhered to in other academic disciplines pertaining to public health, resulting in erroneous but well-publicised claims with broad policy implications before any scientific oversight could be applied.

Download Full-text

A Robust Prediction Model for Species Distribution Using Bagging Ensembles with Deep Neural Networks

Remote Sensing ◽

10.3390/rs13081495 ◽

2021 ◽

Vol 13 (8) ◽

pp. 1495

Author(s):

Jehyeok Rew ◽

Yongjang Cho ◽

Eenjun Hwang

Keyword(s):

Neural Networks ◽

Species Distribution ◽

Best Practice ◽

Deep Neural Networks ◽

Species Distribution Models ◽

Species Distribution Model ◽

Environmental Data ◽

Distribution Model ◽

Distribution Models ◽

Robust Prediction

Species distribution models have been used for various purposes, such as conserving species, discovering potential habitats, and obtaining evolutionary insights by predicting species occurrence. Many statistical and machine-learning-based approaches have been proposed to construct effective species distribution models, but with limited success due to spatial biases in presences and imbalanced presence-absences. We propose a novel species distribution model to address these problems based on bootstrap aggregating (bagging) ensembles of deep neural networks (DNNs). We first generate bootstraps considering presence-absence data on spatial balance to alleviate the bias problem. Then we construct DNNs using environmental data from presence and absence locations, and finally combine these into an ensemble model using three voting methods to improve prediction accuracy. Extensive experiments verified the proposed model’s effectiveness for species in South Korea using crowdsourced observations that have spatial biases. The proposed model achieved more accurate and robust prediction results than the current best practice models.

Download Full-text

Bounding species distribution models

Current Zoology ◽

10.1093/czoolo/57.5.642 ◽

2011 ◽

Vol 57 (5) ◽

pp. 642-647 ◽

Cited By ~ 21

Author(s):

Thomas J. Stohlgren ◽

Catherine S. Jarnevich ◽

Wayne E. Esaias ◽

Jeffrey T. Morisette

Keyword(s):

Species Distribution ◽

Honey Bees ◽

Best Practice ◽

Species Distribution Models ◽

Regression Tree ◽

Classification And Regression Tree ◽

Suitable Habitat ◽

Environmental Predictors ◽

Distribution Models ◽

Africanized Honey Bees

Abstract Species distribution models are increasing in popularity for mapping suitable habitat for species of management concern. Many investigators now recognize that extrapolations of these models with geographic information systems (GIS) might be sensitive to the environmental bounds of the data used in their development, yet there is no recommended best practice for “clamping” model extrapolations. We relied on two commonly used modeling approaches: classification and regression tree (CART) and maximum entropy (Maxent) models, and we tested a simple alteration of the model extrapolations, bounding extrapolations to the maximum and minimum values of primary environmental predictors, to provide a more realistic map of suitable habitat of hybridized Africanized honey bees in the southwestern United States. Findings suggest that multiple models of bounding, and the most conservative bounding of species distribution models, like those presented here, should probably replace the unbounded or loosely bounded techniques currently used.

Download Full-text