Association of Human Mobility and Weather Conditions on Dengue Mosquito in Hong Kong during the COVID-19 Pandemic

DOI: https://doi.org/10.1101/2024.04.17.24306004

Paper Authors

Yufan Zheng, Keqi Yue, Eric W. M. Wong, Hsiang-Yu Yuan

Research Aim

This study aimed to assess the influence of human mobility on the abundance and extensiveness of Aedes albopictus, taking account of the lagged and nonlinear effects of total rainfall and mean temperature based on distributed-lagged non-linear models (DLNM).

Data Description

The data folder contains the following datasets:

Weather Data

Weather data were collected from the Hong Kong Observatory across three regions: Hong Kong Island & Kowloon, New Territories East, and New Territories West (Table 1). The dataset includes daily total rainfall and mean temperature from April 2020 to August 2022. Monthly averages for these metrics were calculated to predict mosquito abundance and extensiveness (Table 2). Weather factors for each region were derived by averaging data from selected weather stations. (https://www.hko.gov.hk/tc/index.html)

Human Mobility Data

Human mobility indices for Hong Kong were sourced from Google, representing behavioral changes and social distancing during the COVID-19 pandemic. Three indices—residential, workplace, and parks—were selected to capture mobility patterns relevant to mosquito-borne disease risk. Indices were calculated relative to a baseline day (median values from Jan 3–Feb 6, 2020) and aggregated monthly for model predictions. (https://www.google.com/covid19/mobility/)

Mosquito Activity Data

Mosquito activity data, provided by the Food and Environmental Hygiene Department, were collected using Gravidtraps since 2020 to monitor Aedes albopictus. (https://www.fehd.gov.hk/english/pestcontrol/risk-pest-mosquito.html)

Table 1: The Monitoring Sites of Mosquito Activity Data in Three Areas in Hong Kong

Area	Site
Hong Kong Island & Kowloon	Chai Wan West, Tin Hau, Shau Kei Wan & Sai Wan Ho, Wan Chai North, Happy Valley, Central, Sheung Wan and Sai Ying Pun, Sai Wan, North Point, Aberdeen and Ap Lei Chau, Pokfulam, Deep Water Bay & Repulse Bay, Cheung Chau, Tung Chung, Tsim Sha Tsui, Mong Kok, Lai Chi Kok, Sham Shui Po East, Cheung Sha Wan, Kowloon City North, Hung Hom, Ho Man Tin, Lok Fu West, Kai Tak North, Wong Tai Sin Central, Diamond Hill, Ngau Chi Wan, Kwun Tong Central, Lam Tin, Kowloon Bay
New Territories East	Tseung Kwan O South, Sai Kung Town, Tseung Kwan, Ma On Shan, Lek Yuen, Yuen Chau Kok, Tai Wai, Wo Che, Tai Po, Fanling, Sheung Shui
New Territories West	Tin Shui Wai, Yuen Kong, Yuen Long Town, Tuen Mun (S), Tuen Mun (N), Tuen Mun West, So Kwun Wat, Tsuen Wan Town, Tsuen Wan West, Ma Wan, Sheung Kwai Chung, Kwai Chung, Lai King, Tsing Yi, Tsing Yi South, Tsing Yi North

Table 2: The Weather Stations in Three Areas of Hong Kong

Areas	Weather Stations (temperature)	Weather Stations (rainfall)
Hong Kong Island & Kowloon	King’s Park, Happy Valley, Wong Chuk Hang	Quarry Bay, Cape D'Aguilar, Happy Valley, King’s Park
New Territories East	Ta Kwu Ling, Sha Tin, Tai Mei Tuk	Ta Kwu Ling, Sha Tin, Tai Mei Tuk
New Territories West	New Tsing Yi Station, Sha Lo Wan, Cheung Chau, Tuen Mun Children and Juvenile Home, Wetland Park	Sha Lo Wan, Cheung Chau, Tuen Mun Children and Juvenile Home, Wetland Park

Folders and Code sources

1. Data: A folder containing the processed data, including mosquito activity, weather data, and human mobility data.

2. Model: A folder containing all fitted models and sensitivity analysis models. In addition, summarizes the fitted model results and sensitivity analysis results.

3. Paper: A folder containing all figures in the paper.

4. results: a folder containing fitted, leave-one-out cross-validation (LOOCV), and projection results.

5. 1_Load_packages_data.R: R script to load data, and prepare the input variables.

6. 2_Model_Construction.R: R script to construct models and record results.

7. 3_Model_Sensibility.R: R script to analyze the model parameter sensibility.

8. 4_LOOCV.R: R script to validate the performance of models by the LOOCV.

9. 5_Standardized.R: R script to standardize the fitted results (output).

10. 6_Projection.R: R script to project the mosquito activity with different human mobility scenarios.

11. 7_Figures_of_paper.R: R script to draw all figures in the paper.

How to Run the Code

Follow these steps to reproduce the results:

1. Load Required Packages and Data:

   • Run 1_Load_packages_data.R to install necessary R packages and prepare input data.

2. Construct Models:

   • Run 2_Model_Construction.R to fit the distributed-lagged non-linear models (DLNM).

3. Sensitivity Analysis:

   • Run 3_Model_Sensibility.R to analyze the sensitivity of the model parameters.

4. Model Validation:

   • Run 4_LOOCV.R to validate model performance using leave-one-out cross-validation.

5. Standardize Results:

   • Run 5_Standardized.R to standardize and prepare results for further analysis.

6. Projection Scenarios:

   • Run 6_Projection.R to generate projections for mosquito activity under various human mobility scenarios.

7. Generate Figures:

   • Run 7_Figures_of_paper.R to create all figures presented in the study.

Notes

• Ensure all required R packages are installed before running the scripts.
• Adjust file paths in the scripts to match your local directory structure if necessary.