Air Pollution per Land Use in the Eastern Caribbean

Abstract

INTRODUCTION: Air pollution has been studied extensively in large and developed nations. However, spatiotemporal trends of pollution in smaller Eastern Caribbean countries have typically been underexplored despite increased emissions driven by development. This study sought to estimate and describe air pollution per specific land use features in the islands of St. Kitts and Nevis where air quality has not been characterized. METHODS: Transport (T), recreation (R) and construction (C) sites were identified using Geographic Information Systems (GIS) and local land use records. In total, 33 monitoring sites were chosen (nT=15, nR=10, nC=8) via stratified random and purposeful sampling. Nitrogen Dioxide (NO2), Ozone (O3) and Particulate Matter (PM) levels were measured actively (1 hour) and passively (2 weeks) in 2015 and 2016. Bivariate analysis, linear mixed effects modeling and land use regression were carried out via SAS 9.4 to evaluate spatiotemporal trends. RESULTS: For active 1-hour sampling, average NO2, O3 and PM10 levels were 29.4 ppb (range: 0-182 ppb), 8.6 ppb (range: 0-106 ppb) and 15.5 µg/m3 (range: 0-302 µg/m3) respectively. Construction NO2 levels were higher than recreation (p<0.01). O3 levels were lower in recreation than construction (p<0.01). Construction PM10 levels exceeded transport (p=0.03). NO2 moderately correlated with O3 (r = 0.54; p<0.01). NO2 was higher in dry season (p<0.05) and near construction (p<0.01); lower in proximity to coast (p=0.01). O3 was inverse to road proximity (p<0.01); higher near construction (p<0.05). PM10 varied by site (p<0.001) and marginally by wind direction (p=0.07). Passive O3 was higher in transport than recreation (p<0.01). DISCUSSION: This study provides evidence of air pollution in small developing nations. Maxima indicate potential for exceedance of World Health Organization guidelines. Land use types, specific land use features, time and meteorology proved significant. Further research is warranted.