Boyin Huang

Global Mean Surface Temperature (GMST) is among the most important indicators of climate change, and its associated uncertainties affect the assessment of historical warming and the formulation of mitigation and adaptation policies. China‐MST 3.0 is a newly updated global surface temperature data set that merges China‐LSAT 2.1 for Land Surface Air Temperature (LSAT) and ERSST v6 for Sea Surface Temperature (SST). In this study, we develop a systematic and traceable uncertainty analysis framework for the construction process of this data set. Specifically, we comprehensively evaluate three components of LSAT uncertainty: observation, analysis, and coverage uncertainties, while describing SST uncertainty in terms of both parametric and reconstruction uncertainties. We also provide a quantitative assessment of the spatial and temporal evolution of these uncertainties. The results show that LSAT uncertainty is generally larger than that of SST and is mainly driven by coverage uncertainty. The overall uncertainty in GMST shows a significant downward trend, with the annual 1 σ uncertainty falling below 0.03°C in recent decades, indicating high data reliability. However, uncertainty was high during the second‐half of the 19th century and remains large at high latitudes in the Southern Hemisphere. Comparative analyses indicate that China‐MST 3.0 is broadly consistent with other data sets in both the magnitude and temporal evolution of GMST uncertainty. These findings demonstrate the utility of China‐MST 3.0 as a valuable tool for evaluating global warming since the 1850s.

Understanding the origins of climate variability (e.g., ENSO and AMV) and their interactions across timescales, as well as assessing model performance in simulating them, relies on robust sea-surface temperature (SST) datasets. Yet, there are numerous instrumental SST products that differ in their bias adjustments and gridding/infilling strategies. These structural choices propagate to key inferences about the climate system such as climate variability indices, the separation of internal and forced components, and teleconnection magnitudes and spatial patterns. Here, we explain why instrumental SST products differ, what these differences imply for climate variability and teleconnection analyses, and which products are best suited for specific applications. We review recent advances in bias adjustment and gridding/infilling of in situ data and assess the implications of these methods for global mean SST evolution and regional variability indices. We find substantial discrepancies in trends during the satellite era among older products, whereas state-of-the-art datasets are much more consistent. State-of-the-art SST datasets are also more consistent with signals from CMIP-class climate models in global mean SST during World War II, Atlantic multidecadal variability indices, and trends in the tropical Pacific zonal gradient — demonstrating the need to carefully choose SST datasets when investigating climate variability. Disagreements persist, however, for early-20th-century warming, which has implications for separating forced response from internal variability, and in data-sparse regions such as the Southern Ocean and Arctic. To support robust, physically interpretable teleconnection diagnostics, we articulate practical principles for dataset selection and highlight the NCAR Climate Data Guide as an evolving resource for updated SST benchmarking.

This study assesses the sensitivity of marine heatwave (MHW) intensity to different MHW thresholds and examines the contributions of linear sea surface temperature (SST) warming and record-breaking SSTs to extreme MHWs (xMHWs) over 2012–24. The xMHW is defined by the high threshold of the 97.5th percentile and is compared to the super MHWs (sMHWs) and MHWs due to the SST warming trend [warming-related MHWs (wMHWs)]. The sMHW is identified when and where the daily SST exceeds the highest historic value during the previous years on a given calendar day. The wMHW is approximated as the difference between xMHW and detrended MHW (dMHW). The dMHW is defined using detrended SSTs at the threshold of the 97.5th percentile. Averaged over 2012–24, sMHW and wMHW account for approximately 52% and 43% of xMHW, indicating a more important role of record-breaking SST than that of SST warming trend. Spatially, sMHW dominates over wMHW in high-latitude oceans and during the period of 2012–18; in contrast, wMHW dominates over sMHW mostly in the tropical oceans and during the period of 2019–24. On global average, sMHW and wMHW increase with time at a rate of 1.1 and 3.3 degree-days yr ⁻¹ . However, their relative contributions to xMHWs diverge: The sMHW ratio decreases at −3% yr ⁻¹ , but the wMHW ratio increases at 3.6% yr ⁻¹ , indicating a growing role of long-term warming in xMHW events. Significance Statement: The extreme marine heatwave (xMHW), super MHW (sMHW), and warming-related MHW (wMHW) in the global oceans have strengthened during the past decade. The sMHW dominated over the wMHW during 2012–18 but was comparable during 2019–24. The strengthening of xMHW mostly resulted from wMHW, highlighting the growing influence of long-term ocean warming.

NOAA’s Extended Reconstructed Sea Surface Temperature (ERSST) is a monthly 2°SST product starting from 1850. Our Part I study indicated that the performance scores of spatial correlation coefficient (SCC) and root-mean-square difference (RMSD) dropped clearly after the mid-1970s in the analysis of ERSST with an artificial neural network (ANN) method. In this Part II study, we demonstrate that ERSST with the ANN method can further be improved progressively in the final ERSSTv6 by the following steps: 1) applying a nearest neighbor check (NNC) quality-control algorithm on ship observations, 2) applying a large-scale (>200 km) filter (LS200) on SST superobservations, and 3) upgrading algorithms in proxy SST from ice concentration. These progressive improvements were assessed against validation and observation datasets. In comparison with ERSST with the ANN method alone, the quality of ERSSTv6 improves in the statistical metrics of SCC and RMSD by 2%–11% and 0.01°–0.24°C, respectively, in the global oceans. In the ice-covered regions, SST bias and RMSD decrease by 0.67° and 0.29°C, respectively. Significance Statement The ERSSTv6 using an artificial neural network (ANN) is further improved by progressively implementing the nearest neighbor check (NNC) in quality control (QC) on ship observations, large-scale filter (LS200) on SST superobservations, and ice-SST proxy algorithm in the ice-covered regions.

NOAA’s Extended Reconstructed Sea Surface Temperature (ERSST) is an operational global SST product based on in situ observations, which has been widely used in monitoring and assessing global ocean climate particularly El Niño–Southern Oscillation (ENSO) events. ERSSTv5 and its predecessors, however, encountered two shortcomings: 1) low SST spatial variabilities in the data-sparse regions before the 1950s and 2) low-performance scores against in situ observations after the 1970s. The first problem has been mitigated in this Part I study of ERSSTv6 by removing a 3-month running average and applying an interpolation method using an artificial neural network (ANN). The improvements of the ANN method over an empirical orthogonal teleconnection (EOT) method used in previous versions were assessed against validation, testing, and observation datasets. In comparison with ERSSTv5, the spatial correlation coefficient (SCC) with reference to observations increases by 5%, and root-mean-square difference (RMSD) with reference to observations decreases by 0.03°C in ERSSTv6. The improvements of SCC and RMSD are more pronounced in the tropical Pacific and the Southern Hemisphere oceans between 60° and 30°S. The second problem has been addressed separately in our Part II study. Significance Statement ERSSTv6 has been upgraded from ERSSTv5 by implementing an interpolation method using an artificial neural network (ANN). In comparison with ERSSTv5, ERSSTv6 has higher spatial coherence and lower root-mean-square error in the global oceans from 1850 to 2021.

NOAA’s Extended Reconstructed Sea Surface Temperature (ERSST) is a monthly 2°SST product starting from 1850. Our Part I study indicated that the performance scores of spatial correlation coefficient (SCC) and root-mean-square difference (RMSD) dropped clearly after the mid-1970s in the analysis of ERSST with an artificial neural network (ANN) method. In this Part II study, we demonstrate that ERSST with the ANN method can further be improved progressively in the final ERSSTv6 by the following steps: 1) applying a nearest neighbor check (NNC) quality-control algorithm on ship observations, 2) applying a large-scale (>200 km) filter (LS200) on SST superobservations, and 3) upgrading algorithms in proxy SST from ice concentration. These progressive improvements were assessed against validation and observation datasets. In comparison with ERSST with the ANN method alone, the quality of ERSSTv6 improves in the statistical metrics of SCC and RMSD by 2%–11% and 0.01°–0.24°C, respectively, in the global oceans. In the ice-covered regions, SST bias and RMSD decrease by 0.67° and 0.29°C, respectively. Significance Statement: The ERSSTv6 using an artificial neural network (ANN) is further improved by progressively implementing the nearest neighbor check (NNC) in quality control (QC) on ship observations, large-scale filter (LS200) on SST superobservations, and ice-SST proxy algorithm in the ice-covered regions.

NOAA’s Extended Reconstructed Sea Surface Temperature (ERSST) is an operational global SST product based on in situ observations, which has been widely used in monitoring and assessing global ocean climate particularly El Niño–Southern Oscillation (ENSO) events. ERSSTv5 and its predecessors, however, encountered two shortcomings: 1) low SST spatial variabilities in the data-sparse regions before the 1950s and 2) low-performance scores against in situ observations after the 1970s. The first problem has been mitigated in this Part I study of ERSSTv6 by removing a 3-month running average and applying an interpolation method using an artificial neural network (ANN). The improvements of the ANN method over an empirical orthogonal teleconnection (EOT) method used in previous versions were assessed against validation, testing, and observation datasets. In comparison with ERSSTv5, the spatial correlation coefficient (SCC) with reference to observations increases by 5%, and root-mean-square difference (RMSD) with reference to observations decreases by 0.03°C in ERSSTv6. The improvements of SCC and RMSD are more pronounced in the tropical Pacific and the Southern Hemisphere oceans between 60° and 30°S. The second problem has been addressed separately in our Part II study. Significance Statement: ERSSTv6 has been upgraded from ERSSTv5 by implementing an interpolation method using an artificial neural network (ANN). In comparison with ERSSTv5, ERSSTv6 has higher spatial coherence and lower root-mean-square error in the global oceans from 1850 to 2021.

In 2023 the NOAA Daily Optimum Interpolation Sea Surface Temperature in the Atlantic set new records, peaking above 24. 3°C for the full basin 30°S–60°N and above 25. 3°C for the North Atlantic 0–60°N, both for the first time during the satellite era. Proposed mechanisms include anomalous radiative and thermodynamic surface fluxes, horizontal transport, changing mixed layer thickness and basal entrainment rates. To test these ideas the heat budget of the ocean mixed layer was examined in a numerical simulation. The results show that the anomalous SSTs in the North Atlantic during the summer of 2023 were caused by increased shortwave warming and reduced thermodynamic cooling. In contrast, ocean mean and eddy heat transport convergence and heat exchanges along the mixed layer base dominate only regionally. A second contributor to the record SSTs was the preconditioning of mixed layer temperature following a series of anomalously warm years.

NOAA Daily Optimum Interpolation Sea Surface Temperature (DOISST) and other similar sea surface temperature (SST) products indicate that the globally averaged SST set a new daily record in March 2023. The record‐high SST in March was immediately broken in April, and new daily records were set again in July and August 2023. The SST anomaly (SSTA) persisted at a record high from mid‐March to the remainder of 2023. Our analysis indicates that the record‐high SSTs, and associated marine heatwaves (MHWs) and even super‐MHWs, are attributed to three factors: (a) a long‐term warming trend, (b) a shift to the warm phase of the multi‐decadal Pacific‐Atlantic‐Arctic (PAA) mode, and (c) the transition from the triple‐dip succession of La Niña events to the 2023–24 El Niño event.