Search for: All records

In this study, we aimed to evaluate the performance of various machine learning (ML) classifiers to predict mode choice of movement-challenged persons (MCPs) based on data collected through a questionnaire survey of 384 respondents in Dhaka, Bangladesh. The mode choice set consisted of CNG-driven auto-rickshaw, bus, walking, motorized rickshaw, and non-motorized rickshaw, which was found as the most prominent mode used by MCPs. Age, sex, income, travel time, and supporting instrument (as an indicator of the level of disability) utilized by MCPs were explored as predictive variables. Results from the different split ratios with 10-fold cross-validation were compared to evaluate model outcomes. A split ratio of 60% demonstrates the optimum accuracy. It was found that Multi-nominal Logistic Regression (MNL), K-Nearest Neighbors (KNN), and Linear Discriminant Analysis (LDA) show higher accuracy for the split ratio of 60%. Overfitting of bus and walking as a travel mode was found as a source of classification error. Travel time was identified as the most important factor influencing the selection of walking, CNG, and rickshaw for MNL, KNN, and LDA. LDA and KNN depict the supporting instrument as a more important factor in mode choice than MNL. The selection of rickshaw as a mode follows a relatively normal probability distribution, while probability distribution is negatively skewed for the other three modes. 

Information and material biological legacies that persist after catastrophic forest disturbance collectively constitute the ecological memory of the system and may strongly influence future stand development. Catastrophic disturbances often result in an influx of coarse woody debris (CWD), and this material legacy may provide beneficial microsites that affect successional and structural developmental pathways. We examined how microenvironmental characteristics influence the regeneration of woody plants in a subtropical woodland that experienced a large influx of CWD from a catastrophic wind disturbance. Specifically, we asked (1) what microenvironmental factors best explain woody plant density, richness, and height in the regeneration layer and (2) does woody plant density, richness, and height benefit from the large influx of CWD to a degree that competition dynamics and succession may be modified? Data were collected in a Pinus palustris woodland that had experienced an EF3 tornado and was subjected to a four-year prescribed fire rotation. We documented live woody plants <5 cm diameter at breast height, soil, and site characteristics and tested for differences in seedling and sapling density, species richness, and height in relation to CWD proximity. We used a random forest machine learning algorithm to examine the influence of microenvironmental conditions on the characteristics of woody plants in the regeneration layer. Woody plant density and species richness were not significantly different by proximity to CWD, but plants near CWD were slightly taller than plants away from CWD. The best predictors of woody plant density, richness, and height were abiotic site characteristics including slope gradient and azimuth, organic matter depth and weight, and soil water content. Results indicated that the regeneration of woody plants in this P. palustris woodland was not strongly influenced by the influx of CWD, but by other biological legacies such as existing root networks and soil characteristics. Our study highlights the need to consider ecological memory in forest management decision-making after catastrophic disturbance. Information and material legacies shape recovery patterns, but, depending on the system, some legacies will be more influential on successional and developmental pathways than others. 

Intense precipitation events (IPE; 99th percentile) in the southeastern United States from 1950 to 2016 were analysed temporally, spatially, and synoptically. The study area was partitioned into latitudinal and physiographic regions to identify subregions that experienced significant changes in IPE frequency or intensity. Furthermore, the spatial synoptic classification (SSC) was used to ascertain what surface weather types are associated with IPEs. Additionally, in conjunction with the SSC, surface forcing mechanisms for the 30 most extreme subregional IPEs were studied to uncover the surface synoptic conditions responsible for IPEs. Results revealed that IPEs increased in frequency and intensity on an annual basis for the southeastern United States. Seasonal results indicated that IPE frequency only increased in the fall. Subregional results reveal that latitudinally, IPEs became more common in the northern latitudes of the study area, while physiographically, significant increases in IPE frequency were most pronounced in areas inland from the Atlantic Coastal Plain. An increase in the annual number of IPEs associated with moist tropical (MT) days was identified across the study area, but was more prevalent in the central and north central latitudinal regions, and areas inland from the Atlantic Coastal Plain outside of the Appalachian Mountains. This MT increase was possibly caused by more common northwards and inland intrusion of these types of IPEs. While moist moderate (MM) and transitional (TR) days were most commonly associated with IPEs, these weather types did not have significant trends. The surface forcing mechanisms most commonly associated with the strongest IPEs were tropical events, followed by stationary fronts and concentric low‐pressure systems. 

Abstract Heat waves have pronounced impacts on human health, ecosystems, and society. Heat waves have become more frequent and intense globally and are likely to intensify further in a warming climate. Across the United States there is a warming trend in average surface temperatures, but concordant increase in heat wave severity appears absent. Limitations in heat waves studies may be responsible for limited detection of a heat wave warming signal. We track daily spatiotemporal evolution of heat waves using geometric concepts and clustering algorithms to investigate how heat manifests on the land surface. We develop a spatial metric combining heat wave frequency, magnitude, duration, and areal extent. We find mixed trends in some individual heat wave characteristics across the United States during 1981–2018. However, exploration of the spatiotemporal evolution of combined heat wave characteristics shows considerable increases during this period and indicates a substantial increase in heat wave hazard across the United States.