The ocean covers more than 70% of our planet and provides life-giving ecosystem services that make life as we know it possible. Nearly 40% of the global population lives within 100km of the coast, a population density that is two times the global average (1). The ocean drives and helps regulates weather and the global climate. It provides food security and livelihood for billions of people, ecosystems services such as absorbing atmospheric CO2 and producing oxygen, genetic resources for the marine biotechnology industry, trade and transport routes, sources of renewable energy, mineral resources, recreation and tourism, and much more. The ocean economy is expected to grow to USD 3 trillion by 2030, employing some 40 million people (2).

To use the ocean sustainably and ensure protection for fragile marine ecosystems, we must monitor and understand how the ocean functions and predict how it will behave in the future as it responds to multiple pressures such as pollution, overfishing and climate change. We measure ocean variables such as temperature, salinity, oxygen, carbon, and chlorophyll to monitor the ocean and ecosystem functioning using global networks of ships, moorings, buoys, and automated underwater vehicles. Together with ocean-monitoring satellites, these data provide a real-time snapshot of ocean conditions.

This snapshot, however, has two limitations: 1) it only offers a glimpse of the ocean in those areas where measurements are available (which are particularly sparse in remote regions such as the Southern Ocean and deep sea); 2) it only captures the status of the ocean over the lifetime of the instrument – it cannot see into the future.
Ocean forecasting is a technique that overcomes these limitations by using algorithms and computer models built on scientific knowledge of how the ocean works to fill in the gaps where data are sparse, to connect a series of these snapshots into a moving picture, and to run that picture forward in time to predict future conditions. This is the same technique used by meteorologists to make weather forecasts that we use in our daily lives.

Forecasting and observing systems work together and in a complementary way. The numerical models need the data from observations for validation of model results and, equally important, to obtain a realistic description of the present conditions in the ocean that serves as starting point for the forecast. This technique for the ‘initialization’ of the model with measured conditions is called ‘data assimilation’ and is vital to obtain realistic forecasts. Conversely, when models fail to reproduce the real ocean, this helps to target where our understanding of ocean processes and ecosystems is weak, requiring additional research and data.

Operational Ocean Forecasting Systems around the world routinely provide ocean forecasts needed for a wide range of activities such as navigation and maritime safety, search and rescue operations, pollution monitoring and response, marine spatial planning and coastal zone management, fishing and aquaculture, and disaster risk reduction and response (hurricanes, tsunamis, storm surges and flooding). The numerical models that are in the core of the ocean forecasting services are also used by scientists to explore global ocean and climate trends such as warming, marine heat waves, sea ice melt, shifting circulation patterns, CO2 uptake, acidification, deoxygenation, and changes to marine ecosystems.