How to Use the Portal?
Main view: Use the menu on the left to view past sea level variability and associated uncertainty across a range of timescales (from months to several years), and the possible tendency towards an increase/decrease in near-future regional sea levels.
On the chart view: You can inspect an individual region where the machine learning model prediction was made. The graph below shows a comparison of satellite mean sea level anomalies with the machine learning modeled / predicted sea level estimates over time. Hover over a point or use the time slider for detailed information. The bar chart allows to visually compare the sea level magnitude over time. Explore the machine learning predictions!
Know What we are Looking For
A note on the context and goals: Our modeled / predicted sea level estimates represent the contribution of temperature changes in the upper layers of open ocean regions (a proxy for climate variability conditions) to future sea-level variations in the coastal areas of these regions. The upper ocean thermal structure is largely influenced by natural internal variability, and it modulates regional sea level variability over periods of several years or decades. Our models offer the opportunity to identify key coastal regions vulnerable to internally induced regional sea level changes across many places around the globe. The modeled (upper ocean temperature-driven) sea-level contribution at specific locations may be more or less than the total sea level observed with satellite due to many local factors (other than internal climate variations) influencing changes in regional sea level, especially on timescales less than 5 years. Complementary analysis of other aspects (e.g., vertical land motion, tidal effects, storms, land ice change) could be performed to help better interpret results where relevant. Read more
Sea Level ML-Modeling Analysis
The methodology used to model and produce the sea level predictions from Nieves et al. (2021), Scientific Reports, www.nature.com/articles/s41598-021-87460-z
Citation: Nieves, V., Radin, C. & Camps-Valls, G. Predicting regional coastal sea level changes with machine learning. Sci Rep 11, 7650 (2021). https://doi.org/10.1038/s41598-021-87460-z
Data Acronyms and What They Mean
95% PI: 95% Prediction Interval.
GP: Gaussian Process regression method.
GP (HO): GP performed on a holdout set.
RNN: Recurrent Neural Network method.