Because of the emission of heat-trapping greenhouse gases by human activities, the natural energy flows have been interfered and currently there is an energy imbalance in the Earth’s climate system. More than 90% of the excess heat is accumulated within the global oceans thus leading to an increase of ocean heat content (OHC). Therefore, OHC is a fundamental indicator of global warming and causes sea level rise via thermal expansion. This presentation will assess the state of knowledge in estimating and understanding OHC changes in recent years. Also, we will provide knowledge gaps and suggested actions to better monitor OHC changes, for example, improve the quantification of the uncertainty in OHC record. We will provide gaps and suggested community actions in understanding the ocean heat uptake and its link to sea surface temperature or global surface temperature, i.e. identifying and addressing the regional hot spots and extremes, better understanding of the temporal variations from seasonal, semi-annual, inter-annual, to decadal scales.

Vertical land motion (VLM) is the connection between absolute sea-level (ASL) from a satellite altimeter and relative sea-level from a tide gauge (TG). VLM is often sparsely observed yet required for understanding and interpreting sea-level rise. Altimeter and TG data have been combined to infer VLM, yet regionally-correlated systematic errors in altimetry are not considered and remain to be fully understood. We develop a Kalman filtering and smoothing framework to simultaneously estimate location-specific VLM and residual mission-specific systematic errors in a common reference frame. We evaluate the performance of the method using along-track data from the Jason-series missions (1992-2020) in the Baltic Sea, where glacial isostatic adjustment is the dominant driver of VLM, and then further apply and extend the method to Australia. In the Baltic Sea region, our approach improves estimates of VLM and ASL at TG locations. We successfully estimate significant local land motion at TGs of up to ~4.5 mm/yr which is otherwise not obtainable through interpolated GPS velocities. Our estimates suggest regional systematic errors within the altimetry can be as large as ~±0.5-2.5 mm/yr, presumably due to regional orbit errors as well as geophysical and environmental mis-modelling. In Australia, we extend the technique to also include data from ERS-2, Envisat, SARAL/AltiKa and Sentinel-3A missions and test our ability to estimate time-dependent VLM and altimeter systematic errors. We estimate that Australia is generally subsiding over the altimeter period but find that any post-seismic deformation or surface loading deformation is below the noise level – noise which appears to be dominated by residual oceanography between the altimeter comparison point and the tide gauge. Overall, our approach advances our ability to estimate local VLM and regional systematic errors using TG and altimeter data and further clarifies uncertainties in such studies.

An improved sea-level budget closure has been found on global mean scale since 1900 and at regional mean scale since the 1960s by the intercomparisions between various reconstructions based on tide-gauge records. However, the local sea-level budget at individual tide gauges has not been critically closed. Here, we consider a hybrid sterodynamic sea-level (SDSL) representing ocean dynamic process, glacial isostatic adjustment (GIA), changes of land ice mass, terrestrial water storage, and residual vertical land motion to evaluate the local sea-level budget at tide gauges over 1958-2015. We find the observed trends at 166 of the total 170 tide-gauge stations distributed globally agree with the sum of contributors within 90% confidence level, with matching mean trend (1.3 mm yr^-1 vs. 1.2 mm yr^-1) and spatial variability (equivalent standard deviation of 2.1 mm yr^-1). The SDSL is the dominant contributor to the trend and spatial variability at tide gauges, except individual stations close to previous melting source where GIA is important.

Before the gravity solutions from the observations of Gravity Recovery And Climate Experiment (GRACE) and GRACE Follow-on (GRACE-FO) satellite missions are used to estimate the Global Ocean Mass Change (GOMC), spatial filtering method must be applied due to the strong noise over the oceans, different filtering methods will cause different signal leakages. Besides signal leakage, many other factors also impact the GOMC estimate, specifically including (1) Restoring the AOD1B background model, (2) Geocenter motion corrections, (3) C20 and C30 corrections, (4) Pole tide corrections, (5) GIA correction, (6) C00 correction, etc. Considering the GRACE gravity solutions from different institutes also caused different GOMC estimates we adopt four recently released GRACE/GRACE-FO models (CSR RL06, GFZ RL06, JPL RL06 and Tongji-Grace2018) to quantitatively investigate the impacts of those factors on GOMC estimates. We confirm that Geocenter motion, C20 and GIA corrections have large impacts on the GOMC estimates, especially for the GIA correction model. Normally the C20, C21, S21 and C30 coefficients are replaced by satellite laser ranging measurements. Note that the AOD1B RL06 model has been contained in the background time variable gravity model when inversing these coefficients. Therefore one should not add the GAD product back again to these replaced coefficients to avoid the ‘‘double counting’’ problem. Besides, if one use the recent GRACE Technical Note 13 product for Geocenter motion correction, whose GIA has been corrected with ICE6G-D, which means that one can only use ICE6G-D for GIA correction to estimate GMOC. Thus it is valuable for analyzing and determining the “best” match GIA model, C20 correction, GC correction, etc, for the most recent GRACE products to improve the accuracy of estimated GOMC. 

Mass changes of land ice (i.e., glaciers and ice sheets) lead to geographically variable patterns in regional sea level, often referred to as “sea level fingerprints”. Current sea level fingerprint products are generally limited to coarse resolutions due to their high computational cost and generally do not consider the uncertainties associated with the underlying ice sheet projections. In this study we use the sea level fingerprint module of the Ice Sheet System Model (ISSM) to provide high-resolution sea level fingerprints for the 21^st century with a particular focus on the uncertainties in fingerprints due to ice sheet projection uncertainty. These are based on large-ensemble simulations of the Antarctic Ice Sheet (AIS) generated by perturbing physical parameters within the Parallel Ice Sheet Model. The ISSM sea level fingerprint module is configured with an unstructured mesh grid, with the flexibility of increasing resolution near melting sources and coasts. Our results show that uncertainty in the description of physical processes within ice sheet models causes large uncertainties in both the global mean sea level contributions and future sea level fingerprints during the 21^st century. Understanding local coastal sea level uncertainty requires an understanding of uncertainty at drainage basin level in ice sheet projections, not just the uncertainty of the entire ice sheet.

Global sea level is rising primarily because global temperatures are rising, causing ocean water to expand and land ice to melt. However, sea-level rise is not uniform; it varies from place to place. Singapore and Southeast Asia show significant variability that depends on the combination of global mean sea-level rise and regional factors, such as ocean and atmospheric circulation patterns, the gravitational and deformational effects of land ice mass changes, and tectonic vertical land motion. The relative influence of these regional factors determines whether rates of local sea-level change are higher or lower than the global mean, and by how much. Understanding the physical processes driving global and regional changes is key to predicting the impact of rising seas and extreme events. Determining the rates, mechanisms and geographic variability of sea-level change is a priority science question for the next decade of ocean research. The SouthEast Asia SEA-level program (SEA2) will integrate instrumental, historical and geological sea-level datasets in Southeast Asia with sophisticated modeling capabilities to improve the accuracy of projections of sea-level rise and extreme sea levels, and to communicate the results to the scientific community, governmental agencies and the public. SEA2 will assemble a multi-disciplinary team of leading experts in the fields of reconstructing past and present sea-level change, polar ice-sheet history, oceanography, geodesy, glacial isostatic adjustment (GIA) modeling, and the statistical analysis and modeling of sea-level data. Through training, the SEA2 team will build a home-grown scientific community that can respond to Singapore and Southeast Asia’s need for future sea-level projections and their interpretation.

Sea level is changing due to climactic factors, and tidal range may also change due to sea level change. To understand how tidal behavior (range) changes with the sea level changes, we analyzed the freely available hourly sea-level data, GESLA (Woodworth, 2017), focused on the coast of Southeast Asian Sea where a wide range of tidal behaviour and tidal ranges have been shown. This observed sea-level is a combined effect of a tide component and a storm surge component. Therefore we analysed the sea-level data separately for understanding the causation of the sea level change. We used the tidal harmonic and the correlation analysis methods to quantify the tidal evolution related to the sea level changes. Here we focused along the Malay Peninsula and showed how tidal evolution e.g. the form factor calculated as (K1+O1)/(M2+S2), MSL (mean sea level), have been changed temporally and spatially. From this result, we assume that the sea level around this area has been changed mainly by the tidal behavior which has changed by the surge. This conclusion is confirmed by the tide-only and the surge simulations using a NEMO numerical model. The water density driven baroclinic effects were not considered in this study.