Publications

Bonafoni, S; Keeratikasikorn, C (2018). Land Surface Temperature and Urban Density: Multiyear Modeling and Relationship Analysis Using MODIS and Landsat Data. REMOTE SENSING, 10(9), 1471.

Abstract
This work aims to model and relate the urban density and land surface temperature (LST) by a straightforward and efficient approach. Although the urban density-LST relation is widely addressed in literature, this study allows for its modeling and parameterization in an accurate way, providing a further scientific support for the city planning policy. The urban density and the LST analysis is carried out in the Bangkok area for the years 2004, 2008, 2012, and 2016; in this time interval, the city exhibited an evident urban expansion. Firstly, by using land cover maps obtained from Landsat reflective observations, the urban land density growth across the years studied is evaluated by applying a ring-based approach, a method employed in urban theory, providing urban density curves as a function of the distance from the city center. For each year, the urban density curve is well modeled by an inverse S-shape function, the parameters of which highlight an urban sprawl over the years studied and an outskirt growth in recent years. Then, employing 237 MODIS LST images, the night-time and daytime mean LST patterns for each year were processed applying the same ring-based analysis, obtaining LST trends versus distance. Albeit the mean LST decreases away from the city core, the daytime and night-time trends are different in both shape and values. The daytime LST exhibits a trend also modeled by an inverse S-shape function, whereas the night-time one is modeled by a quadratic function. Finally, the urban density-LST relationship is inferred across the years: For daytime, the relation is quadratic with a coefficient of determination r(2) around 0.98-0.99, whereas for night-time the relation is linear with r(2) of the order of 0.95-0.96. The proposed approach allows for reliable modeling and to straightforwardly infer a very accurate urban density-LST relationship.

DOI:
10.3390/rs10091471

ISSN:
2072-4292