Floods can severely impact the economy, environment and society. These impacts can be direct and indirect. Past research has focused more on the former impacts. Of the indirect impacts, those on mold growth in indoor environments that affect human respiratory health (e.g. asthma) have received limited attention. Models can be used to predict these impacts and support development of mitigation and preventive actions. Despite the presence of models for some other impacts of flooding, quantitative models for estimating the impacts of flooding on indoor mold spores are lacking. In this article, we studied the aftermath of two recent hurricanes-Ida and Ian-in the United States and applied machine learning algorithms to develop the first quantitative model for predicting mold spores in buildings. A comprehensive fine-scale database (building level), consisting of flood characteristics, building properties, human indoor activities and existing mold spores, prepared through survey questionnaires, home inspections, laboratory analyses and flood hindcasting, from 60 homes was utilized. The modeling results suggested satisfactory performance for regression-based predictions of indoor mold spores (coefficient of determination or R2 of 0.83 and 0.38). This is the first quantitative model for predicting the impacts of flooding on mold spores. Our study provides a foundation for quantitative assessments of flood impacts on indoor mold spores in residential buildings. This supports insurance companies, public health officials and emergency managers to better assess the impacts of hurricanes and extreme flooding on human respiratory health.
In the absence of adequate observations on the off‐channel areas, flood models are typically trained and validated against stream water depths. This approach can be efficient for physics‐based models, which incorporate the underlying physical processes, but the efficiency for data‐driven models like machine learning (ML) algorithms is unclear. The existing off‐channel observations like high‐water marks (HWMs) are also subject to uncertainty. This paper addressed three research questions: (a) how useful are ML models, trained with stream gauges, for hindcasting water depths in the off‐channel areas? (b) how does incorporating the uncertainty of HWMs improve the model performance? and (c) does the uncertainty incorporation improve the model transferability to other watersheds and events? To answer these questions, we evaluated the performance of ML models across three large coastal watersheds in the US during three hurricanes—Michael, Ida and Ian. The model was developed under three scenarios, which differed in terms of the flood observational data (stream gauges and HWMs) used for their training and validation. A loss function was proposed to incorporate the uncertainty of observations. We found that ML models trained solely by stream gauges performed well only for stream hindcasts. Satisfactory hindcasts on off‐channel areas were obtained by incorporating the HWMs' uncertainty via the loss function. This uncertainty incorporation reduced the model bias and resulted in the best transferability to other coastal watersheds and flood events. Our study provides insights about developing transferable ML models for hindcasting water depths on streams and off‐channel areas in coastal watersheds during extreme events.