PubAg

Main content area

An analysis on the urban heat island effect using radiosonde profiles and Landsat imagery with ground meteorological data in South Florida

Author:
Kandel, Hari, Melesse, Assefa, Whitman, Dean
Source:
International journal of remote sensing 2016 v.37 no.10 pp. 2313-2337
ISSN:
1366-5901
Subject:
Landsat, drainage, heat island, meteorological data, rangelands, remote sensing, summer, surface temperature, thematic maps, Florida
Abstract:
Decades-long effects of drainage and development in south Florida that began in early to mid-20 ᵗʰ century have resulted in the loss of natural forested and rangelands, and expansion of agricultural and urban areas. This has brought a change in the thermalscape of the surrounding areas. Surface and atmospheric surface-layers become drier and warmer in anthropogenic land covers compared to natural ones, introducing an effect called the urban heat island. This study aims to analyse the spatial and temporal existence of the urban heat island on land surface and near-surface atmosphere during summer using land-use/land-cover data, surface-based weather station records, radiosonde profiles, and Landsat-5 Thematic Mapper images captured between 1970 and 2012 in south Florida. Urban cover increased by about 10% from 1974 to 2011. The station-based urban–rural temperature difference (T ᵤ₋ᵣ is about 4°C over the entire area of south Florida and Landsat-derived surface temperature difference is about 3°C in the Landsat-covered area. They both show temporal increase throughout the period at a rate higher than the rate of increase of global average temperature. A decreasing near-surface diurnal temperature range of about ‒1°C (p = 0.005) and increased lifting condensation level (>20 m) were detected from the Miami radiosonde. Satisfactory validation results of surface and near-surface temperature (Nash–Sutcliffe coefficient = 0.70, coefficient of determination = 0.79) from the eastern urban stations further substantiate the findings on urban heat islands explored by the observed data.
Agid:
5243547