Integrated Forecasts Based on Public Health Surveillance and Meteorological Data Predict West Nile Virus in a High-Risk Region of North America

Abstract

Background: West Nile virus (WNV), a global arbovirus, is the most prevalent mosquito-transmitted infection in the United States. Forecasts of WNV risk during the upcoming transmission season could provide the basis for targeted mosquito control and disease prevention efforts. We developed the Arbovirus Mapping and Prediction (ArboMAP) WNV forecasting system and used it in South Dakota from 2016 to 2019. This study reports a post hoc forecast validation and model comparison. Objectives: Our objective was to validate historical predictions of WNV cases with independent data that were not used for model calibration. We tested the hypothesis that predictive models based on mosquito surveillance data combined with meteorological variables were more accurate than models based on mosquito or meteorological data alone. Methods: The ArboMAP system incorporated models that predicted the weekly probability of observing one or more human WNV cases in each county. We compared alternative models with different predictors including a ) a baseline model based only on historical WNV cases, b ) mosquito models based on seasonal patterns of infection rates, c ) environmental models based on lagged meteorological variables, including temperature and vapor pressure deficit, d ) combined models with mosquito infection rates and lagged meteorological variables, and e ) ensembles of two or more combined models. During the WNV season, models were calibrated using data from previous years and weekly predictions were made using data from the current year. Forecasts were compared with observed cases to calculate the area under the receiver operating characteristic curve (AUC) and other metrics of spatial and temporal prediction error. Results: Mosquito and environmental models outperformed the baseline model that included county-level averages and seasonal trends of WNV cases. Combined models were more accurate than models based only on meteorological or mosquito infection variables. The most accurate model was a simple ensemble mean of the two best combined models. Forecast accuracy increased rapidly from early June through early July and was stable thereafter, with a maximum AUC of 0.85. The model predictions captured the seasonal pattern of WNV as well as year-to-year variation in case numbers and the geographic pattern of cases. Discussion: The predictions reached maximum accuracy early enough in the WNV season to allow public health responses before the peak of human cases in August. This early warning is necessary because other indicators of WNV risk, including early reports of human cases and mosquito abundance, are poor predictors of case numbers later in the season. https://doi.org/10.1289/EHP10287

Multi‐Model Prediction of West Nile Virus Neuroinvasive Disease With Machine Learning for Identification of Important Regional Climatic Drivers

“…Additionally, the top climate covariates in the RF and NN models were not consistent with those selected for the AR(1)-Climate models. While the exogenous covariates for the AR(1)-Climate models were selected from a somewhat different, yet overlapping set of candidate variables, we would expect to see similar types (i.e., temperature or precipitation anomalies) selected across models due to the attributed importance of climate to accurate WNV prediction (DeFelice et al, 2018;Hahn et al, 2015;Hess et al, 2018;Landesman et al, 2007;Shaman et al, 2010Shaman et al, , 2011Wimberly et al, 2022). In a few instances, similar climate anomalies (i.e., temperature or precipitation) were identified by the AR(1)-Climate and machine learning models, but often with different lags.…”

“…In contrast, lagged historical cases up to 5 years often had consistently high importance (rank and relative magnitude). Previous studies using distributed lags (up to 36 months) of meteorological conditions (e.g., temperature, precipitation, and drought) similarly found significant lagged effects on WNND cases at the county scale, but did not consider lags up to 5 years (Davis et al, 2018;Smith et al, 2020;Wimberly et al, 2022). However, none of these models were compared to a historical baseline, like a NB model, to assess relative skill.…”

“…Thus, while climate factors can modulate WNV transmission, climate variables are not the only driver and the relative influence of climate likely can change between years and locations due to the multi‐faceted ecology of WNV. Previous improvements in WNV prediction with climate have been achieved on smaller spatial and temporal scales (Davis et al., 2018; DeFelice et al., 2018; Hess et al., 2018; Keyel et al., 2019; Wimberly et al., 2022). We attempted to develop local forecasts within climate regions (i.e., 200 km diameter hexagons) where other ecological factors are likely more consistent, but WNV dynamics may be consistent only at even smaller geographical scales necessitating even more flexibility.…”

A Spatially Resolved and Environmentally Informed Forecast Model of West Nile Virus in Coachella Valley, California

“…WNV transmission is not only dependent on factors such as bird immunity, and mosquito feeding behavior but is also substantially driven by environmental drivers such as meteorological and hydrological conditions (Davis et al., 2017; DeFelice et al., 2017; Kilpatrick et al., 2006; Paull et al., 2017; Shaman et al., 2005; Wimberly et al., 2022). Much effort has been made to produce accurate forecast models for WNV transmission, however, there remains significant variability and little consensus between these products (Barker, 2019; DeFelice et al., 2017; Keyel et al., 2021; Little et al., 2016; Wimberly et al., 2022). Temperature, humidity, and available water affect the development and survival of WNV mosquito vectors, as well as the extrinsic incubation period (EIP) of the virus (Epstein, 2001; Reisen et al., 2008; Shaman et al., 2005; Wegbreit & Reisen, 2000).…”

“…WNV transmission is not only dependent on factors such as bird immunity, and mosquito feeding behavior but is also substantially driven by environmental drivers such as meteorological and hydrological conditions (Davis et al, 2017;DeFelice et al, 2017;Kilpatrick et al, 2006;Paull et al, 2017;Wimberly et al, 2022). Much effort has been made to produce accurate forecast models for WNV transmission, however, there remains significant variability and little consensus between these products (Barker, 2019;DeFelice et al, 2017;Keyel et al, 2021;Little et al, 2016;Wimberly et al, 2022). Temperature, humidity, and available water affect the development and survival of WNV mosquito vectors, as well as the extrinsic incubation period (EIP) of the virus (Epstein, 2001;Reisen et al, 2008;Wegbreit & Reisen, 2000).…”