Geometry and adjacency effects in urban land surface temperature retrieval from high-spatial-resolution thermal infrared images

Highlights

• An UCM-RT model taking into account geometry and adjacency effects was deduced.

• Accuracy of Urban TES algorithm based on urban effective emissivity is about 0.8 K.

• The magnitude of geometry and adjacency effects on urban LST is up to 2.1 K.

• Application of high spatial resolution TIR image in urban LST retrieval was analyzed.

Abstract

Urban land surface temperature (ULST) is a key surface feature parameter in urban heat island studies. However, the geometry and adjacency effects are usually neglected in conventional land surface temperature (LST) retrieval methods. In this study, considering the geometry and adjacency effects, a new urban canopy multiple-scattering thermal radiative transfer (UCM-RT) model and an urban temperature-emissivity separation (UTES) algorithm were developed for ULST retrieval from Chinese GaoFen-5 (GF-5) satellite thermal infrared (TIR) images. The UCM-RT model and the UTES algorithm incorporate multiple scattering within the urban building canopy and consider the thermal radiance contribution from adjacent pixels and the atmosphere. Two schemes were designed to evaluate the geometry and adjacency effects quantitatively in this study: (1) evaluating their impact on ground temperature and top of atmosphere (TOA) brightness temperature for GF-5 four TIR bands from the simulation dataset, and (2) comparing the retrieval temperature difference between the UTES algorithm and conventional temperature-emissivity separation (TES) algorithm (without considering geometry and adjacency effects). For a specific simulation situation, the magnitude of the geometric effects on urban ground temperatures were found to be approximately 3.2 K, 4.2 K, 4.0 K, and 3.9 K for the four GF-5 TIR bands, while the effects on the TOA brightness temperature reached 2.4 K, 3.3 K, 3.3 K, and 3.0 K for those bands, respectively. The results show that the geometry effects have a significant influence on the thermal radiative transfer process. The temperature error of the UTES algorithm was generally up to 0.8 K, much lower than that of the conventional TES algorithm (~2.2 K) in urban surface. The results indicate that the UTES algorithm is suitable for ULST retrieval, and the retrieval temperature difference between the UTES algorithm and conventional TES algorithm ranges from 0.1 K to 2.1 K, especially for dense building pixels. Additionally, the thermal radiation from adjacent pixels in high-spatial-resolution TIR images has significant influence on ULST retrieval. The comparison between the ULST algorithm for urban land surface and the conventional TES algorithm for the flat surface indicated that the three-dimensional structure inside the pixel has a significant influence on the ULST retrieval. The findings of this study indicate that the geometry and adjacency effects must be considered in ULST retrieval for highly accurate results.

Introduction

Land surface temperature (LST) is a physical variable belonging to the set of Essential Climate Variables (ECVs) required by the United Nations Framework Convention on Climate Change (UNFCCC) and the World Climate Organization (WCO) to have an impact on grand societal challenges, including the United Nations Sustainable Development Goals (SDGs) (Bojinski et al., 2014; UN, 2021). With the increase in buildings and impervious surfaces, urbanization has changed the original natural surface coverage (Weng, 2012; Zhu et al., 2019). Urban surfaces with built-up areas are more effective in absorbing energy than horizontal surfaces owing to the complex urban geometry, which reduces the direct reflection to the atmosphere and increases multiple-scattering (Offerle et al., 2003). Urban geometry includes building height, building density, height/width ratio and sky view factor (SVF) (Brown et al., 2001; Grimmond, 2007; Sharmin et al., 2017). These urban characteristics alter the natural surface energy and radiation balances such that the urban land surface temperature (ULST) is relatively high. The phenomenon of urban heat islands (UHIs) then arises, with a significant temperature difference between urban and rural areas (Manoli et al., 2019; Oke, 1982; Voogt and Oke, 2003). The ULST is a critical parameter for studying the UHI effect (Lai et al., 2018; Weng, 2009). With the development of thermal infrared (TIR) remote sensing techniques, the ULST obtained from remotely sensed data provides a synoptic view and an effective measure of the UHI, by observing the temperature differences between urban surfaces and the surrounding rural regions (Bektaş Balçik, 2014; Weng, 2009; Zhou et al., 2018).

Current LST retrieval algorithms always assume that the pixel is flat, uniform, and spatially isothermal (Li et al., 2013). This assumption is used to ensure that the LST is invariant among different bands and does not cause much error for most natural surfaces. However, urban pixels are much more complicated and often consist of various components with different temperatures and emissivities (Voogt, 1995; Zhan et al., 2013). Owing to the three-dimensional structure inside and around the pixel, the multiple-scattering radiance cannot be ignored in the modeling of pixel thermal emissions. Yang et al. (2015b) analyzed the geometry effect on the ULST and specifically demonstrated that such an effect can cause a temperature difference of approximately 2 K, and this difference is increased with the decrease in SVF. The directional anisotropy of the ULST caused by different geometries was analyzed, and the results showed that the effective anisotropy should be considered in ULST retrieval (Lagouarde et al., 2012; Voogt and Oke, 1998; Zhan et al., 2013). The heterogeneity of the urban surface causes a considerable number of mixed pixels owing to the generally coarse spatial resolution of the TIR remote sensing data. The methods for separating the component temperatures through directional TIR data in vegetation and soil areas have been studied; however, these methods are usually only valid for vegetated surfaces (Jia et al., 2003; Ren et al., 2013). Currently, ULST retrieval algorithms that consider the geometric effects and material characteristics are rare, primarily because of the complexity of geometry inside the urban area, which requires further study.

Directional simulations of surface temperatures in urban areas have been studied (Lagouarde et al., 2012; Lagouarde et al., 2010; Wang et al., 2018a; Wang et al., 2020; Wang et al., 2018b) however, these simulation models require many input parameters and are usually difficult to apply when retrieving large-area urban surface temperatures. Hence, the thermal radiative transfer process considering the geometric characteristics within urban areas and its application in remote sensing retrieval are critical for ULST retrieval studies. Some studies started from the process of thermal radiative transmission and used single-band TIR data or split window algorithms to retrieve the ULST (Yang et al., 2015b; Yang et al., 2016). Currently, there is no ULST retrieval algorithm for the complex underlying surface of an urban area that is suitable for three or more TIR channels. However, more infrared band information is helpful in improving the accuracy of LST retrieval (Li et al., 2013; Nie et al., 2020). The temperature emissivity separation (TES) algorithm with stability and good accuracy is widely used in LST retrieval, for example, the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) LST products (AST_08) (Gillespie et al., 1998; Sabol et al., 2009). However, as we previously mentioned, the traditional TES algorithm regards the urban surface as a flat surface, similar to the other LST retrieval algorithms. Yang et al. (2016) compared the ASTER LST products with the ULST with the geometry effect and found that the conventional TES is not suitable for ULST retrieval, because the empirical relationship between the minimum spectral emissivity and the maximum minimum differences was not established for three-dimensional urban surfaces. Thus, it is necessary to establish a new relationship for high-density urban areas.

In this study, the TIR data from the Visible and Infrared Multiple-Spectral imager (VIMS) on the Chinese GaoFen-5 (GF-5) satellite were used to investigate the impact of the SVF on the ground temperature and top of atmosphere (TOA) brightness temperature and the ULST retrieval algorithm development. The GF-5 satellite was launched on May 9, 2018 and contains four TIR bands (8.01–8.39 μm, 8.42–8.83 μm, 10.30–11.30 μm, and 11.40–12.50 μm) with a spatial resolution of 40 m, much finer than ASTER (90 m), Landsat 8 (100 m) and MODIS (1 km). Several papers have reported LST retrieval from VIMS images using the split-window algorithm, the TES algorithm, and a hybrid algorithm combining the two (Chen et al., 2019; Ren et al., 2018; Yang et al., 2018; Ye et al., 2017), but all of them focused on the flat surface, and therefore cannot be directly applied to estimate the ULST. Nevertheless, these studies provided us with some references to develop the specified ULST algorithm. Because the TES algorithm is available for at least three TIR bands and has been used in some TIR sensors, the primary aim of this study is to refine the TES algorithm for urban pixels based on a new thermal radiative transfer model and definitions.