Water Deeply recently talked to the study’s lead author, Bryan Mundhenk, a PhD candidate in atmospheric science.
Water Deeply: What is our ability to predict atmospheric rivers today? What are the challenges?
Bryan Mundhenk: In general, the current generation of numerical weather prediction models provides very useful information about atmospheric rivers a week or two out. Some of the decision tools – like those available [from the Center for Western Weather and Extremes] at Scripps – are based on these models to help anticipate atmospheric river timing and intensity along the U.S. West Coast.
We focused this study on the subseasonal timescale, beyond the point where today’s numerical models generally lose skill. We defined this subseasonal timescale as forecast lead times [that are] out two to five weeks. A lot of resource decisions are made within this time scale – think about adjusting reservoir levels, anticipating energy needs or planning outdoor events – but no tools are available to guide decision-makers in regions prone to atmospheric river landfall.
Water Deeply: What are the MJO and QBO?
Mundhenk: The Madden-Julian oscillation, or MJO, is the dominant mode of intraseasonal variability in the tropics. It represents large clusters of storms in the tropics that progress along the equator. The MJO is quite variable, but these storm clusters wrap around the globe every 30–90 days.
The quasi-biennial oscillation, or QBO, characterizes the winds in the stratosphere high above the tropics. Our interest in the QBO is based on research from a few years ago that revealed that the state of these upper-level winds can influence the MJO, as far as the strength of the storms in the tropics and the speed at which they propagate around the globe [goes].
Water Deeply: How do the MJO and QBO influence atmospheric rivers?
Mundhenk: This concept is exploiting the ability of the atmosphere to support teleconnections. That is, when and where weather in one location on the globe is related to – and can influence – weather at a distant location.
The tropical storms characterized by the MJO can elicit a wave response that travels around the globe. Think of a storm in the tropics as a stone dropped into a pond. The resulting ripples in that pond are like the wavelike teleconnection response to the stone being dropped. Granted, the ripples in the pond dissipate quickly, but the large-scale response to tropical forcing can take weeks to impact distant parts of the globe.
Water Deeply: It sounds like you’ve developed not a weather model, but a new method of observing and interpreting the MJO and QBO. Is that right?
Mundhenk: That’s right. This method was really an attempt to see if the teleconnection responses could provide useful information within the subseasonal time scale. We used the initial state of the MJO and the QBO to predict anomalous atmospheric river activity two to five weeks in the future.
For example, say on a day in the winter there are strong storms in the Indian Ocean – we call this MJO phase 1 – and easterly winds in the stratosphere over the tropics. Then one should expect higher than normal likelihood of landfalling atmospheric river activity near coastal British Columbia approximately three weeks in the future.
Water Deeply: How accurate are your predictions using this method?
Mundhenk: In that example about heightened atmospheric river activity near British Columbia three weeks into the future, the prediction of increased activity would be correct a maximum of 20 times out of 30. Perhaps more important than the skill value itself, at this stage, is that this finding suggests the dynamics of the atmosphere – the teleconnection responses – can provide useful information about atmospheric river activity well into the subseasonal time scale.
Water Deeply: Do your predictions also provide accuracy about where atmosphere rivers will make landfall?
Mundhenk: We applied this technique to a few regions along the West Coast of North America, but the regions were fairly expansive. It is important to note that this technique targets periods of anomalous atmospheric river activity, not individual atmospheric rivers themselves.
Many locations along the West Coast have pronounced wet seasons during which atmospheric rivers are more common. Say, for a given region, one atmospheric river makes landfall approximately every five days during winter. That would be the “normal.” This technique attempts to predict deviations from that normal.