Comparison of Spatial Interpolation and Regression Analysis Models for an Estimation of Monthly Near Surface Air Temperature in China

et al. 2019

JGR Atmospheres

Self Cite

The single‐channel (SC) algorithm has been widely used to retrieve land surface temperature from Landsat series data for its simplicity and requirement of only one thermal infrared channel. The main error sources of the existing SC algorithms are the linearization of the Planck's function and atmospheric correction. This paper proposed a practical SC (PSC) algorithm to retrieve land surface temperature from Landsat series data aiming at avoiding the aforementioned error sources. The sensitivity of the PSC algorithm to the input parameters was analyzed. The performance of the proposed PSC algorithm was compared with the most commonly used SC algorithm (the generalized SC, GSC) using a simulation data set and satellite measurements. Results showed that the PSC algorithm was less sensitive to uncertainties in the input parameters than the GSC algorithm. When validated with the simulation data set, the root‐mean‐square error (RMSE) of the PSC algorithm was 1.23 K, with an improvement by 0.57 K compared with the GSC algorithm. For the validation with 71 clear‐sky Landsat 8 images, the RMSE of the PSC algorithm was 1.77 K when using the measurements from U.S. surface radiation budget network as real values. Compared with the GSC algorithm, the RMSE improvement for the PSC algorithm was 0.47 K. We conclude that the PSC algorithm is more accurate than the GSC algorithm and the sensitivity to input parameters in the PSC algorithm is weaker than in the GSC algorithm.

Section: Discussionmentioning

confidence: 99%

Section: Methods and Datamentioning

confidence: 99%

A Practical Single‐Channel Algorithm for Land Surface Temperature Retrieval: Application to Landsat Series Data

Wang

et al. 2019

JGR Atmospheres

Self Cite

“…However, these aspects could not thorough respect urbanization of China's six megacities. More key indicators (i.e., population density, air temperature [47], food production [48], cropland losses [14], etc.) of cities and megacities should be investigated in the future.…”

Section: Megacity Terrain Referencementioning

confidence: 99%

Assessing Interannual Urbanization of China’s Six Megacities Since 2000

Liu

et al. 2019

Remote Sensing

As a large and populous developing country, China has entered the rapid urbanization stage since 2000. Until 2018, China has accounted for nearly 1/5 of global megacities. Understanding their urbanization processes is of great significance. Given the deficiencies of existing research, this study explored the interannual urbanization process of China's six megacities during 2000-2018 from four aspects, namely, the basic characteristics of urban land expansion, expansion types, cotemporary evolution of urban land-population-economy, and urbanization effects on the local environment. Results indicated that (1) urban lands in China's six megacities increased by 153.27%, with distinct differences across megacities; (2) all of six megacities experienced the expansion processes from high-speed to low-speed, but they varied greatly in detail; (3) the speeds of urban land expansion in China's megacities outpaced the population growth but lagged behind in GDP increase; and (4) urbanization has triggered an environmental crisis, which is represented by the decline in vegetation coverage and the increase in land surface temperature in newly expanded urban lands. This study enriched the content of urbanization, supplemented the existing materials of megacities, and provided a scientific reference for designing rational urban planning.Despite conducting denser social activities and causing more drastic urban sprawl effects than other medium-and small-sized cities [9], megacities must face more serious eco-environment problems, such as air pollution [10,11], flood disaster [12,13], and cropland losses [14], which have been proven to reflect the urbanization characteristics of a country or region better [15].Urbanization hotspots are expected to emerge in developing countries in the next several decades of the 21st century [1]. As a large and populous developing country with nearly 1/5 of the megacities worldwide, China has experienced significant urbanization at an increasing rate of 232.48% since the implementation of the Reform and Opening-up Policy in 1978 and the country has entered into the rapid expansion stage since 2000 [16]. Therefore, China is a good research area for investigating the urbanization process. During the past decades, urban expansion in China has been thoroughly investigated by using remote sensing technology. These studies have identified the following features. (1) Research objects ranged from a single city (i.e., Beijing [5], Shanghai [17], Harbin [18], Guangzhou [19], Nanjing [20], etc.) to urban aggregations (i.e., Beijing-Tianjin-Hebei [21], Pearl River Delta [22], Yangtze River Delta [23], etc.). However, limited investigations have been conducted on China's six megacities.(2) Research approaches included but not limited to various classification models (i.e., supervised or unsupervised classification, artificial neural network, etc.) and analysis methods (i.e., spatiotemporal dynamics, sprawl patterns, etc.). However, urbanization is a comprehensive process involving both physical morphology and ...

“…Chen et al achieved RMSE values of 2.32 K for Tmax and 2.61 K Tmin using geographically weighted regression (GWR) [15]. Wang et al compared different methods in estimating monthly Ta and found that GWR has a similar prediction performance to that of Kriging regression [35]. A RMSE of 1.8 K for estimating monthly mean Ta was obtained using machine learning algorithms with a set of inputs [27].…”

Section: Cross Validationmentioning

confidence: 99%

“…Statistical models are the most widely used approach to estimating Ta due to its simplicity and interpretability. Many studies have applied different types of statistical models including ordinary regression [7,16,[28][29][30], step-wise linear regression [31], mixed effects regression [32][33][34], geographically weighted regression [15,35,36], regression Kriging [35,[37][38][39], and the hierarchical Bayesian space-time model [40].…”

Section: Introductionmentioning

confidence: 99%

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

2019

Remote Sensing

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.