A Bayesian Kriging Regression Method to Estimate Air Temperature Using Remote Sensing Data

Abstract

Surface air temperature (Ta) is an important physical quantity, usually measured at ground weather station networks. Measured Ta data is inadequate to characterize the complex spatial patterns of Ta field due to low density and unevenness of the networks. Remote sensing can provide satellite imagery with large scale spatial coverage and fine resolution. Estimating spatially continuous Ta by integrating ground measurements and satellite data is an active research area. A variety of methods have been proposed and applied in this area. However, the existing studies primarily focused on daily Ta and failed to quantify uncertainties in model parameter and estimated results. In this paper, a Bayesian Kriging regression (BKR) method is proposed to model and estimate monthly Ta using satellite-derived land surface temperature (LST) as the only input. The BKR is a spatial statistical model with the capacity to quantify uncertainties via Bayesian inference. The BKR method was applied to estimate monthly maximum air temperature (Tmax) and minimum air temperature (Tmin) over the conterminous United States in 2015. An exploratory analysis shows a strong relationship between LST and Ta at the monthly scale, indicating LST has the great potential to estimate monthly Ta. 10-fold cross-validation approach was adopted to compare the predictive performance of the BKR method with the linear regression method over the whole region and the urban areas of the contiguous United States. For the whole region, the results show that the BKR method achieves a competitively better performance with averaged RMSE values 1 . 23 K for Tmax and 1 . 20 K for Tmin, which are also lower than previous studies on estimation of monthly Ta. In the urban areas, the cross-validation demonstrates similar results with averaged RMSE values 1 . 21 K for Tmax and 1 . 27 K for Tmin. Posterior samples for model parameters and estimated Ta were obtained and used to analyze uncertainties in the model parameters and estimated Ta. The BKR method provides a promising way to estimate Ta with competitively predictive performance and to quantify model uncertainties at the same time.

Urban Heat Island and Its Regional Impacts Using Remotely Sensed Thermal Data—A Review of Recent Developments and Methodology

“…Classification with LST, index, albedo Miles and Esau [63], Trlica et al [64], Bonafoni [65], Wong and Nichol [66], Jin [67], Wu et al [68], and Hu and Brunsell [69] Regression models, geostatistical analysis Zhang and Du [70], Wicki and Parlow [71], Dai et al [72], Song et al [73], Sellers et al [74], Du et al [75], Shahraiyni et al [76], Chun and Guldmann [77], Ho et al [78], and Lai et al [79] Multiple sensors, data fusion Huang and Wang [80], Li et al [81], Berger et al [82], Liu et al [83], Fu and Weng [84], Liang and Weng [85], and Dousset and Gourmelon [86] Machine learning, decision support information system Chakraborty and Lee [87], Mpakairia and Muvengwi [88], Zhang et al [89], Tran et al [90], Shahraiyni et al [76], Weng and Fu [91], Mallick et al [92], Connors et al [93], Wentz et al [94], Xian and Crane [95], Wilson et al [96], and Xian et al [97] [98]. Generally, an urban heat island (UHI) is an urban area or metropolitan area that is significantly warmer than its surrounding rural areas because of human activities.…”

Spatiotemporal Distribution Characteristics of Reference Evapotranspiration in Shandong Province from 1980 to 2019

“…Interpolation methods have significant advantages in analyzing the spatial distribution characteristics and future trends of climate. Radial basis function (RBF) [22], kriging [23], Bayesian kriging regression (BKR) [24], inverse distance weight (IDW), and other interpolation methods have been widely used. Among them, IDW is a spatial distribution method that fully considers the regional relationship between various factors.…”

Compréhension du microclimat urbain lyonnais par l’intégration de prédicteurs complémentaires à différentes échelles dans des modèles de régression

“…Plus récemment, la modélisation par régression (Boer et al, 2001) ou encore par les réseaux de neurones et les techniques d'apprentissage de maching learning (Antonić et al, 2001) est apparue. De plus, de multiples études ont abordé cette question, par ces interpolations spatiales classiques déterministes (Wang et al, 2017) ou stochastiques (Zhang et Du 2019) ou bien par ces régressions multiples (Cantat, 2004 ;Carrega et Rosa, 2005 ;Cristobal et al, 2006 ;Hengl et al, 2012 ;Zhu et al, 2013 ;Chen et al, 2016 ;Kastendeuch et al, 2016 ;Carrega et Martin, 2017 ;Mira et al, 2017 ;Richard et al, 2017 ;.…”