Published 16:54 IST, September 3rd 2020
NASA researchers use machine-learning to better predict hurricane intensity
NASA scientists have used machine learning to develop an experimental computer model that promises to greatly improve accuracy of hurricane intensity detection
Advertisement
Accurately predicting wher a hurricane will undergo rapid intensification – where wind speeds increase by 35 mph (56 kph) or more within 24 hours – is incredibly difficult. But researchers led by scientists at NASA's Jet Propulsion Laboratory have used machine learning to develop an experimental computer model that promises to greatly improve accuracy of detecting rapid-intensification events.detec
Eyeing inner workings
re are two parts to a hurricane forecast: its track and its intensity. Scientists and forecasters have gotten very good at predicting where a hurricane will make landfall. But forecasting its strength still gives m trouble because it depends on surrounding environment as well as what's happening inside se storms. Properties such as how hard it's raining or how quickly air is moving vertically are challenging to measure inside a hurricane.
Advertisement
It's also difficult to determine which internal characteristics result in rapid intensification of se storms. But after sifting through years of satellite data, researchers found that a good indicator of how a hurricane's strength will change over next 24 hours is rainfall rate inside storm's inner core – area within a 62-mile (100-kilometer) radius of eyewall, or dense wall of thunderstorms surrounding eye. harder it's raining inside a hurricane, more likely storm is to intensify. team gared this rainfall data from Tropical Rainfall Measuring Mission, a joint satellite project between NASA and Japanese Aero Exploration ncy that operated from 1997 to 2015.
In addition, researchers found that changes in storm intensity depended on ice water content of clouds within a hurricane – measurements y gared from NASA's CloudSat observations. temperature of air flowing away from eye at top of hurricanes, kwn as outflow temperature, also factored into intensity changes. Researchers obtained outflow temperature measurements from NASA's Microwave Limb Sounder (MLS) on Aura satellite as well as from or datasets.
Advertisement
(Part of NASA's fleet of wear- and climate-tracking satellites, CloudSat uses advanced radar to examine inner structure of clouds, helping researchers better understand how severe tropical cyclones, as well as climate changes related to clouds, occur. IM: NASA)
Advertisement
Using computational algorithm capabilities
team added rainfall rate, ice water content, and outflow-temperature predictors to ones US National Hurricane Center already uses in its operational model to come up with ir own predictions via machine learning. re are so many variables inside a hurricane, and y interact in such complex ways, that many current computer models have trouble accurately depicting inner workings of se storms.
Advertisement
Machine learning, however, is better able to analyze se complex internal dynamics and identify which properties could drive a sudden jump in hurricane intensity. researchers used computational algorithm capabilities of IBM Watson Studio to develop ir machine learning model.
n y trained ir model on storms from 1998 to 2008 and tested it using a different set of storms, from 2009 to 2014. Researchers also compared performance of ir model with Center's operational forecast model for same storms from 2009 to 2014.
Advertisement
For hurricanes whose winds increased by at least 35 mph (56 kph) within 24 hours, researchers' model had a 60% higher probability of detecting rapid-intensification event compared to current operational forecast model. But for those hurricanes with winds that jumped by at least 40 mph (64 kph) within 24 hours, new model outperformed operational one at detecting se events by 200%.
Researchers are w testing ir model on storms during current hurricane season in US to gauge its performance. In future, y plan to sift through satellite data to find additional hurricane characteristics that could improve ir machine learning model. Predictors such as wher it's raining harder in one part of a hurricane versus ar could give scientists a better look at how storm's intensity might change over time.
"It's an important forecast to get right because of potential for harm to people and property," said Hui Su, an atmospheric scientist at JPL part of research team.
16:54 IST, September 3rd 2020