Estimating Outdoor and Indoor Temperature Exposure for the BiSC Participants

Extreme heat events pose a significant public health risk and are expected to become more frequent and severe as climate change accelerates. Pregnant women are particularly vulnerable to extreme heat, which can have harmful effects on both maternal health and fetal development. To better understand these risks in the BiSC, we estimated outdoor and indoor temperature exposure for our participants throughout pregnancy to assess their impact on health outcomes in the future.

To estimate outdoor temperature, we used a Random Forest model that integrates data from multiple sources, including the outdoor temperature data from meteorological stations, hourly skin* and 2m temperature**, total precipitation, wind components, the height of the land, and greenspace data. [ref] This approach allowed us to generate an accurate temperature estimate map (R² = 0.98) at a fine spatial (250m) and daily temporal resolution. During the study period, BiSC participants experienced the average daily outdoor temperatures with a mean of 17.28°C, a maximum of 30.1°C in summer, and a minimum of 4.59°C in winter.

The indoor temperatures were monitored over the course of one week during both the first and third trimesters. Indoor temperatures were generally higher than outdoor, with a mean of 21.5°C, a maximum of 33.88°C in summer, and a minimum of 12.42°C in winter. Then, we developed an ensemble model*** using more than 50 predictor variables to predict the indoor temperature throughout the whole pregnancy. This model incorporated meteorological conditions (e.g., outdoor temperature and solar radiation), home characteristics (e.g., the floor, type of floor finish and wall, orientation of windows, and volume of the bedroom), socio-demographic factors (e.g., maternal age and number occupants), and occupant behaviors (e.g., window opening habits). Most of the predictor variables were collected by participant questionnaires and home visits conducted by our field workers.

These temperature exposure estimates will play a crucial role in evaluating the potential effects of heat exposure on health during pregnancy. By understanding how both indoor and outdoor temperatures influence maternal and fetal health, we can contribute to strategies aimed at reducing risks associated with extreme heat events.

We sincerely thank all BiSC participants for their valuable contributions to this research!

 

* Hourly skin: temperature taken from the surface of the floor, considered as the “skin” of the Earth.

** 2m temperature: temperature of the air measured from 2 meters above the surface.

*** Ensemble model: machine learning model that combines several individual models to improve performance, as in the case of techniques such as Random Forest.

 

This news has been written by Yu Zhao, a pre-doctoral student of the BiSC Project.