Observational Assessment of Changes in Earth's Energy Imbalance Since 2000

Abstract

Satellite observations from the Clouds and the Earth's Radiant Energy System show that Earth's energy imbalance has doubled from 0.5 ± 0.2 Wm^-2 during the first 10 years of this century to 1.0 ± 0.2 Wm^{- 2} during the past decade. The increase is the result of a 0.9 ± 0.3 Wm^-2 increase absorbed solar radiation (ASR) that is partially offset by a 0.4 ± 0.25 Wm^-2 increase in outgoing longwave radiation (OLR). Despite marked differences in ASR and OLR trends during the hiatus (2000-2010), transition-to-El Niño (2010-2016) and post-El Niño (2016-2022) periods, trends in net top-of-atmosphere flux (NET) remain within 0.1 Wm^-2 per decade of one another, implying a steady acceleration of climate warming. Northern and southern hemisphere trends in NET are consistent to 0.06 ± 0.31 Wm^-2 per decade due to a compensation between weak ASR and OLR hemispheric trend differences of opposite sign. We find that large decreases in stratocumulus and middle clouds over the sub-tropics and decreases in low and middle clouds at mid-latitudes are the primary reasons for increasing ASR trends in the northern hemisphere (NH). These changes are especially large over the eastern and northern Pacific Ocean, and coincide with large increases in sea-surface temperature (SST). The decrease in cloud fraction and higher SSTs over the NH sub-tropics lead to a significant increase in OLR from cloud-free regions, which partially compensate for the NH ASR increase. Decreases in middle cloud reflection and a weaker reduction in low-cloud reflection account for the increase in ASR in the southern hemisphere, while OLR changes are weak. Changes in cloud cover in response to SST increases imply a feedback to climate change yet a contribution from radiative forcing or internal variability cannot be ruled out.

Keywords: Climate change; Clouds; Earth radiation budget; Earth’s energy imbalance; Satellite.

Fig. 1

Cloud fraction by cloud class for September 2002. Global coverages of each cloud class are as indicated

Fig. 2

Global mean all-sky TOA flux anomalies and multivariate ENSO index (MEI) from CERES EBAF Ed4.2 for 03/2000–12/2022. a ASR and –OLR; b NET

Fig. 3

Zonal mean all-sky TOA flux trends for 03/2000–12/2022. a ASR; b –OLR; c NET

Fig. 4

Regional trends in a ASR, b –OLR, c NET (Wm⁻² per decade), and d SST (K per decade) for 03/2000–12/2022. Hatching indicates trends significant at 2.5–97.5% confidence level. SSTs are from ECMWF Reanalysis 5 (ERA5) (Hersbach et al. 2020)

Fig. 5

Twelve-month running average global anomalies in ERA5 SST and CERES a ASR, b OLR (positive up, since –OLR is displayed with a reversed y-axis), and c NET TOA radiation. Period considered: 03/2000–12/2022

Fig. 6

Monthly time series a, c and trends b, d for MEI (top) and anomalies in ERA5 SST (bottom). White circles in b and d correspond to trends that exceed the 2.5–97.5% CI. Time period 03/2000–12/2022

Fig. 7

Trends in solar irradiance (SOL), –SW, ASR, –OLR, and NET TOA flux for a all-sky, b clear-sky and c CRE. White circles indicate trends that exceed the 2.5–97.5% CI

Fig. 8

Trends in all-sky and clear-sky flux, CRE, clear fraction weighted clear-sky column (Clear Wtd) and cloud fraction weighted cloudy column (Cloud Wtd) flux contributions for –SW, –OLR (–LW), and NET TOA flux from FBCT product. Error bars correspond to 2.5–97.5% CI. Time period: 07/2002–12/2022

Fig. 9

Contribution to zonal mean –SW trend from a clear-sky, b low cloud, c middle cloud, d high cloud, e polar cloud, f all. Period considered: 07/2002–12/2022. The SH and NH hemispheric average trends for each cloud type are indicated in each figure

Fig. 10

Same as Fig. 9 but for clear-sky and cloud fraction

Fig. 11

Zonal low-cloud trends with contribution from Cu, SCT and Sc. Period considered: 07/2002–12/2022

Fig. 12

Same as Fig. 9 but for –OLR

Fig. 13

Clear-sky frequency and cloud fraction trends by cloud type from: a MODIS for NH, b CC for NH, c MODIS for SH, d CC for SH, e MODIS for globe, and f CC for globe using coincident measurements from 01/2008 to 12/2017