I have been studying the water cycle over the oceans since the 1980s. At first, it was a neglected research area; who cared if it rained on the ocean? But I knew that the water cycle was the driver for salinity variations in the ocean – which affect ocean density, and thus ocean currents and mixing – nearly as much as temperature variations. When we compute the sums of how much water is evaporated from the ocean, and how much of it rains back out onto the ocean, we realize that the oceanic water cycle completely dominates the global water cycle. Indeed, the return flow of all rivers in the terrestrial water cycle amounts to less than 10% of the water evaporated from the ocean.

Yet, if you Google image search “Global Water Cycle” you will find most results show only a small corner with a tiny bit of ocean, while most of the image is devoted to terrestrial processes. Of course, we live on land and are highly dependent on that portion of the water cycle, so it is an understandable bias – but it seems to have led to a consistent misperception of how the water cycle actually works. We live on the blue planet, 70% covered by water, and to focus only on the 30% area covered by land will give a distorted impression of the water cycle. In particular, terrestrial hydrologists tend to focus on the local recycling of water by evaporation from lakes and rivers and evapo-transpiration from plants, and they ignore the significant transport of water from ocean to land.

For me, the importance of such water transfers came into sharp relief in 1993, when the upper Midwest experienced exceptional rainfall and massive flooding on the Missouri, Mississippi and Illinois rivers. A huge pulse of water poured down the Mississippi, overflowing the banks and flooding St. Louis, Memphis and New Orleans successively through the summer. Next, the extra water entered the Gulf of Mexico and caused a significant decrease in upper ocean salinity.Oceanographers from Texas and Florida documented the movement of the strong low salinity anomaly into the Loop Current and into the Gulf Stream. It struck me then that if all that flood water was freshening the ocean, sometime prior to the floods there must have been a significant increase in ocean salinity as part of the supply of all that rain. That is, no watershed floods by recycling its own water; the extra water had to come from the ocean.

Conservation of water and salt on the planet guarantees that some part of the ocean had to get saltier to supply the rain that flooded the Midwest that summer. Because the atmosphere can only hold a few centimeters of water at a time, I also knew that the areas of net export of water from the ocean tend to be comparable to areas of net import of water. In effect, the subtropical high-pressure systems in mid-latitudes (where evaporation exceeds precipitation) are comparable in area to the regions where precipitation exceeds evaporation – the high latitude oceans, the intertropical convergence zones, and over land.

The subtropics are the sunniest regions on Earth, where the dry, descending air of the Hadley cells (a large-scale atmospheric convection cell in which air rises at the equator and sinks at mid-latitudes) extracts water and latent heat energy from the ocean solar collector. These are also the saltiest regions of the open ocean. So, I knew that if a number of states experienced an extra foot of rainfall over a given month, some comparable area of ocean had to export an extra foot of water by evaporation, and would be noticeably saltier than normal. That is, the processes of evaporation and precipitation directly affect Sea Surface Salinity (SSS), and the SSS anomalies are only slowly mixed away by weak ocean mixing and transport processes. Thus, I began looking to SSS variations to predict floods and droughts on land. If some part of the ocean gets saltier than normal, then we can be sure that some other part of the world will experience more rainfall. Similarly, if an area of ocean is fresher than normal, then we know that some areas will experience less rain than usual.

As a “soft-money” researcher it proved surprisingly difficult to secure funding to develop this unconventional idea. Eventually, I was able to convince a talented rainfall expert, Dr. Laifang Li, who had a WHOI Postdoctoral fellowship, to take on the project. She quickly hit paydirt, finding that spring SSS anomalies in the eastern North Atlantic were a good predictor of summer rain in Africa, while SSS anomalies in the western North Atlantic were good predictors of summer rain in the US Midwest. The three-month lead time was a surprise, since atmospheric moisture is only held for about 10 days. She was able to show that an interesting interplay between soil moisture and the patterns of atmospheric winds could explain the long lead time. Moreover, SSS was a much better predictor of terrestrial rainfall than was sea surface temperature or any of the various modes of atmospheric variation, such as ENSO or NAO. A diagram of the process is shown in the figure above.

Salient Predictions is using new insights (like these) in the inherent predictability of ocean properties for weather on land to provide a new type of long-range weather forecast. So, what are we predicting for the Midwest in the summer of 2020? The image above shows the March 2020 salinity anomalies' effect on rainfall patterns around the US. We are seeing that the Gulf of Mexico is exceptionally salty, and much of the western North Atlantic is high in salinity, too. The Gulf is also exceptionally warm this spring, feeding lots of moisture and latent heat energy into the Southeastern US, which is already experiencing a very stormy spring. The latent heat released when atmospheric moisture condenses into rain is a main driver of the atmospheric convection, making it essential to thunderstorms, tornadoes and hurricanes – we use the SSS as an indicator of that export of latent heat from the oceans.

At Salient, we feel confident saying that the Midwest will experience a wetter-than-normal summer. The high salinities off the East coast and the high soil moisture in the South, especially, are indicators of a wet summer for the Midwest. Our precipitation forecast for June (from mid-April) is shown below, as a percent of normal. While the South should return to near-normal conditions, the Northern plains and upper Midwest should experience above-normal rainfall in early summer.

If you are a farmer or agricultural vendor/service provider selecting seeds or fertilizers for the coming growing season, or a hydropower operator anticipating the state of your flow and reservoirs, please reach out to Salient Predictions to leverage our skill in long-range forecasts of future weather conditions.