Assessment of the Response of Land Surface Temperature to Land Use and Land Cover in Kano Metropolis and its Suburbs

Abstract

Kano State is faced with increasing air and surface temperature caused as a result of the continuous development activities, constructions and influx of people to the state. The study therefore analyses the development of land surface temperature (LST) on different land cover and land use categories and to determine the possible impact of the various class on LST. Landsat 8 Operational Land Imager and Thermal Infrared Sensor (OLI/TIR), multi-temporal remote sensing satellite data of 2015 and 2016 were used to retrieve LST and derive land use and land cover classification map using random forest machine learning algorithm and various land cover indices such as Normalised Difference Vegetation Index (NDVI), Normalised Difference Built Index (NDBI) and Modified Normalised Difference Water Index (MNDWI) were derived using R statistics software. These land cover indices were used to examine the landscape attributes, characteristics and to further understand the cause-effect relationship between LST and LULC using a Pearson’s correlation analysis and simple linear regression model. LULC classification map showed that using several multi-temporal satellite imageries for classification to extract biophysical information provide a more accurate result with a kappa coefficient of 1.017 and 1.013 and overall accuracy of above 85% which showed an excellent agreement between the map and ground truth data. The retrieved LST pointed out that land surface temperature could be as high as 38oC to 40oC in hot seasons and as low as 22oC to 25oC in wet seasons. LST values were extracted for the different land cover and land use class and result revealed that there is a decreasing trend of LST all through the season from built up areas (such as residential, commercial and industrial) which recorded a higher LST to water bodies (such as lakes, ponds, streams and rivers) which showed a low LST value. The correlation analysis generated between LST and the three land cover indices showed that for all time steps MNDWI showed a negative correlation with LST (-0.313 to -0.686 and -0.208 to -0.786 in 2015 and 2016 respectively). Likewise, NDVI showed a higher negative correlation of between -0.127 to -0.436 and -0.137 to -0.389 in 2015 and 2016 respectively. While, NDBI revealed a high positive correlation with LST of between 0.491 to 0.804 and 0.666 to 0.839 in 2015 and 2016 respectively. Urban Heat Island (UHI) effect was described by determining hot and cold spots areas with the core of the study area characterised to be hotspot areas while the periphery and most notably the western part of the study area where irrigation fed agriculture are practiced characterised to be cold spot areas and this explains why there is a decreasing trend of surface temperature as one move from the core to the periphery. This study infer that vegetation plays a vital role in weakening LST and recommend that tree planting campaign should be carried out, landscaping should be done alongside road or bridge constructions, urban greening concept should be carried out by town planners and individuals to reduce the effect of UHI.