Recent progress on evaluating and analysing surface radiation and energy budget datasets

ABSTRACT

Although the surface energy budget is essential to determine Earth’s climate, site measurements of various radiative components are still too scarce to properly characterize their spatial and temporal variations. This has led to the development of a growing number of surface radiation products, mainly including remotely sensed data, model reanalysis data, and simulations using General Circulation Models (GCMs). This collection of papers introduces new techniques, including the use of machine learning methods for radiation estimation, and evaluates and compares various radiation products, as well as their spatio-temporal variations. These studies show large discrepancies among various products across nearly all radiative parameters in either accuracy or spatio-temporal variations. However, remotely sensed radiation products perform relatively better than others. Despite this, there is an urgent need for further efforts to address these discrepancies and improve the accuracy of these estimates. Even though the major radiative parameters including downward shortwave radiation, net longwave radiation, and albedo, from most products show insignificant long-term variation trends on a global scale, only specific regions, such as the Yunnan-Kweichow Plateau (YKP) and regions with permafrost (i.e. Qinghai-Tibet Plateau and Arctic) and glaciers (i.e. Altai Mountains) exhibit remarkable trends.

KEYWORDS:

• Surface radiation
• energy budget
• estimation
• evaluation
• spatio-temporal variation

1. Introduction

The surface radiation energy budget governs the surface energy balance and is represented by the net radiation (R_n), which characterizes the balance between incoming radiation from the atmosphere and outgoing radiation from the Earth’s surface. Mathematically, all-wave net radiation is the sum of shortwave and longwave net radiation (R_ns / R_nl) and is expressed by: (1) ${\mathrm{R}}_n=R_{ns}+R_{nl}=\left(1-\alpha \right)R_s^d+R_l^d-R_l^u$(1) where $R_s^d$ is the downward shortwave radiation, $R_l^d$ and $R_l^u$ are the downward and upward longwave radiations, respectively, and $\alpha$ is the land surface broadband albedo defined as the ratio of the reflected shortwave radiation to the incoming shortwave radiation. The downward direction is usually defined as positive.

The R_n is essential to characterize the land surface ecological, hydrological, and biogeochemical processes (Liang et al. 2019) and its partitioning can be written as: (2) $R_n=G+H+\lambda ET$(2) where G is the soil heat flux, H is the sensible heat flux, and λET is the latent heat flux (LE) in which λ is the latent heat of evaporation of water and ET is the rate of evaporation of water.

From a global perspective, the Earth’s energy budget represented by R_n at the top-of-atmosphere (TOA) (R_n,TOA) is referred to as the balance between the energy coming into the Earth system from the Sun at the TOA (R_so) and the energy lost to space in the forms of reflected shortwave radiation ($R_{s,TOA}^d$) and emitted longwave radiative fluxes ($R_{l,TOA}^u$) from the Earth’s surface and atmosphere: (3) ${\mathrm{R}}_{n,TOA}=\left(1-{\alpha }_{TOA}\right)R_{so}-R_{l,TOA}^u$(3) where ${\alpha }_{TOA}$ is the planetary albedo, defined as the fraction of incident global mean shortwave radiative flux reflected to space.

Therefore, radiative studies mainly focus on the following components: all-sky $R_s^d$, albedo at the surface ($\alpha$) and at the TOA (${\alpha }_{TOA}$), longwave radiation at the surface ($R_l^d$,$R_l^u$, and R_nl), all-wave all-sky net radiation at the surface (R_n) and at the TOA (R_n,TOA), and latent (LE) and sensible heat (H) fluxes.

With technological advances in remote sensing, modern numerical weather modeling, and assimilation, an unprecedented wealth of datasets relevant to surface radiation and the energy budget has become available over the past several decades. The collection named “Comprehensive Evaluation and Spatio-temporal Analysis of Surface Radiation and Energy Budget Datasets” was launched in March 2022 to explore the discrepancies among the radiative datasets at various tempo-spatial scales, the sources of uncertainties in their magnitudes and mechanisms, the spatio-temporal variations in different surface energy budgets and the corresponding driving factors, as well as new progress in estimation methods. This collection has gathered 12 papers regarding new developments and studies for modeling, product evaluation, and spatio-temporal variations in terms of various radiative parameters. As Guest Editors, we have taken the opportunity to receive, read, and learn about novel and interesting contributions to this wide topic. A brief summary of the papers in this collection is given in the next section.

2. Overview of the articles featured in this collection

All of the papers adhere closely to the collection subject and can be roughly divided into three major topics: methods, product evaluation, and spatio-temporal variation analysis.

Three papers are about methods for radiation estimation, two of which introduce new techniques and one compares nine machine learning approaches. Specifically, Li et al. (2023) improved a method for estimating surface shortwave radiation in East Asia from MODIS data by optimizing the calculation of cloud-related parameters (i.e. cloud transmittance and reflectance). Zhan and Liang (2023) presented a new model to estimate $R_l^u$ at 1 km by directly linking MODIS TOA radiances with the $R_l^u$ determined by CERES (Clouds and the Earth’s Radiant Energy System) and other information with machine learning methods. Li et al. (2022) developed and evaluated nine machine learning methods, including six classic ML methods and three deep learning (DL) methods, applied to the estimation of surface daily R_n directly from MODIS TOA observations. These papers reveal that machine learning methods, particularly DL methods, are becoming increasingly popular in the field of radiation estimation with satellite data; additionally, the use of information from other sources combined with satellite data, such as reanalysis data and in situ measurements, could more effectively improve estimation accuracy than using only satellite data.

The evaluation of various radiative products is always a popular topic, especially for product producers and a wide range of users. In this collection, six papers introduced their evaluations on different radiative products, including remotely sensed products (i.e. CERES, GLASS, and so on), reanalysis products (i.e. ERA5, MERRA2, and so on), and GCM simulations, in terms of the comprehensive radiative parameters including $R_s^d$, albedo, longwave radiation, R_n, and latent and heated flux. Yang et al. (2023) and Tong et al. (2023) evaluated seven and six representative $R_s^d$products over global and China, respectively. They both found that most products have the tendency to overestimate and that satellite retrievals generally show relatively better accuracy than others. Some of them (i.e. GLASS, BESS, and MCD18A1) could meet the threshold accuracy requirement suggested by the World Meteorological Organization for the application of global numerical weather prediction (20 W/m²) at monthly scales. However, the accuracy of all products is distributed unevenly in space, such as the larger uncertainties usually appeared within the regions at high-/low- latitudes and subjected to strong human activities (e.g. urbanization). Large discrepancies exist among the multi-year mean $R_s^d$ of these products (Yang et al. 2023) over the globe, both over land and oceans. Similarly, studies of longwave radiation and R_n products conducted by Xu, Liang, and Jiang (2022) and Yin et al. (2023) also concluded that the remotely sensed products, particularly the newly released GLASS series product, tend to perform better than other kinds of products, except ERA5, but it usually has a larger value. Discrepancies among products varied spatially and temporally. In terms of albedo, Liu, Qiang, and Ying (2023) found that most products show good consistency at the global scale, especially after 2000, but differences among them are more significant in the low- and high- latitudes. In addition, Ma et al. (2022) evaluated the performance of LE and H simulations from 30 CMIP6 models in the permafrost region over the Qinghai-Tibet Plateau (QTP) and the Arctic. The authors concluded that the simulations performed better in the Arctic than in the QTP, but the accuracies of the simulations of the two parameters were influenced by various factors over the two study regions.

The spatio-temporal variation in radiative parameters was also a popular research topic. Compared to other parameters, there were more studies on shortwave radiation balance ($R_s^d$ and albedo) variations. For example, Yang et al. (2023) analyzed global long-term variation trends of $R_s^d$ estimates from various products and they found that different $R_s^d$products show different trends, although most trends are not significant; Cheng et al. (2023) analyzed the $R_s^d$trends and the driving factors over the Yunnan-Kweichow Plateau (YKP) based on the GLASS product, and indicated that the aerosol optical thickness (AOT), El Niño/La Niña, and the monsoon could explain the decreased $R_s^d$ over the YKP, especially the AOT. Yue et al. (2022) investigated the spatio-temporal variations in albedo and their linkages with the mass balance of the Muz Taw Glacier in the Altai Mountains using the MODIS albedo product and found a decreased trend in May and a V-shaped seasonal variability of albedo from May to September. Additionally, variations in the two parameters were speculated from the changes of other parameters in some studies, such as the declined $R_s^d$ and a slight decrease in albedo over China for 2003–2018 found by Jiang et al. (2023) through the reduced potential evapotranspiration (PET), and the increased $R_s^d$ over the permafrost region of the QTP and the Arctic found by Ma et al. (2022) through the rapid increase in H. Regarding the other radiative parameters, Xu et al. (2023) observed an increased trend in $R_l^d$and $R_l^u$ between 1982 and 2015, followed by a decreased trend during 2016–2021 for GLASS-AVHRR and ERA5, respectively, but the net longwave radiation did not exhibit a significant trend. Yin et al. (2023) indicated that no significant long-term trends in R_n were observed over the globe, but the variations of various products were remarkably different, especially before 2000. Overall, most studies pointed out that the existing radiative products (i.e. $R_s^d$ and R_n) may not be suitable for long-term analysis.

3. Conclusions and innovation opportunities

The work presented in this collection remarks on the significant role that remote sensing technologies play in retrieving and analyzing the surface radiation and energy budget. However, it is acknowledged that a large discrepancy exists among the available radiative products, which are far from meeting the requirements of various applications, especially in terms of accuracy and spatio-temporal resolutions. To improve the satellite-based radiative estimates with higher accuracy and finer spatio-temporal resolution, new models that combine remotely sensed data and machine learning algorithms are very promising, particularly DL algorithms, which are more appropriate for big data; however, the way to effectively integrate physical or complex mechanisms into these models is critical to further improve their performance. Secondly, the statistical methods (e.g. Liu, Qiang, and Ying (2023)), the distribution, size, and processing scheme of the site measurements used for validation, and the spatio-temporal scales considered (e.g. Tong et al. (2023)) significantly influence the final results; hence, future work requires a careful design of the evaluation scheme for various radiative parameters. The performance and spatio-temporal variations of various radiative estimates in extreme regions, such as high-elevation areas, rugged terrains, and polar regions have yet to be thoroughly explored. Hence, the primary emphasis remains on acquiring data, either from ground measurements or from high-quality products. In addition, other essential radiative parameters that usually receive little attention, such as the radiative fluxes at the TOA and the direct and diffuse solar radiation, should be given more weight in the future. Overall, we hope that the work presented in this collection will inspire innovative concepts and methods that will improve our understanding and estimation of various radiative parameters.

Acknowledgements

We would like to thank the authors who contributed to this collection and the reviewers for their helpful suggestions and criticisms that improved the papers in the collection. We also want to extend our gratitude to Professor Changlin Wang and Ms. Tricia for their editorial support and guidance in preparing the collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).