Startup Strives to Take Risk Out of Severe-Weather Predictions

TECHNOLOGY: Statistical Approach Focuses on Extreme Heat, Cold Waves

By Andrea Siedsma

Monday, July 19, 2010

Scientists and meteorologists expect that severe-weather events such as the recent heat wave in the Midwest and Northeast will become more frequent as the climate continues to change. What if harsh weather conditions such as heat and cold waves could be predicted well in advance to give energy companies and consumers a head start on managing their resources? 

photo

Stephen Bennett

EarthRisk Technologies, a San Diego-based startup, plans to do just that. EarthRisk was born out of technology created by the Scripps Partnership for Hazards and Environmental Applied Research, or SPHEAR, at UC San Diego’s Scripps Institution of Oceanography. SPHEAR recently created new methods to predict severe cold weather with a lead time of up to 40 days, which could be used to develop decision-support software that links weather and climate research to develop custom risk assessments for supply chain management and energy consumption.

“There is a lot of data out there in the one- to two-week time frame for weather prediction, but once you go beyond day 15 the amount of information available drops dramatically,” said Stephen Bennett, co-founder of EarthRisk and director of business development for Scripps Institution of Oceanography. “Weather is a highly complex thing that is forced by so many variables. It’s hard to precisely model it by creating a mini atmosphere in a computer model, which is the primary approach for today’s weather predictions. So what we have done is create algorithms based on statistical techniques that use past weather as a predictor for future weather. The idea is to augment existing modeling technologies with a probabilistic statistical approach focused on extremes.”

Quantifying Historic Events

To create this model, Bennett’s colleagues at Scripps, associate research scientist Alexander Gershunov and postdoctoral scholar Kristen Guirguis, designed a process that starts by quantifying historic weather events. That data is then algorithmically used to link preceding weather conditions at locations around the globe to future outbreaks of extreme weather. The researchers used six weather variables, including jet stream position, air temperature and thunderstorm activity in the Indian Ocean, to develop the process. The research results are undergoing scientific peer review and have been submitted for publication in various academic journals. UCSD has also filed copyright and patent protection to enable commercial use of the research outcomes.

“We’re not trying to give a traditional weather forecast 30 days in advance,” Bennett said. “What we have developed is a tool that helps predict when the most severe cold outbreaks tend to occur. When we see certain weather conditions at various locations on the planet, we also tend to observe that outbreaks of cold weather in the Midwest U.S. tend to follow at various lag times. The jet stream over Alaska is an important pattern, for example. When a certain jet stream pattern evolves over Alaska, about 10 to 15 days later 70 to 80 percent of the time a severe-cold outbreak occurs in the Midwest.”

EarthRisk has raised nearly $500,000 in capital, Bennett said, including some funding from San Diego-based venture capital firm SEAR Technologies and from a $50,000 grant through the Clean Tech Innovation and Commercialization Program. Under the program – a collaborative effort among the City of San Diego, the von Liebig Center at the UCSD Jacobs School of Engineering, and San Diego State University, with funding from private companies – Bennett and his colleagues will receive $50,000 for proof of concept funding and mentoring assistance from business advisers at the von Liebig Center to help commercialize their technology. Such funding and support, Bennett said, has been critical for EarthRisk in securing outside funding.

“The von Liebig Center’s program guides faculty and their students throughout the process of commercialization and helps demonstrate the commercial viability of their discoveries,” said von Liebig director Rosibel Ochoa. “It contributes to de-risking the technology, therefore making it more attractive for an investor or entrepreneur to partner with the innovators.”

Bennett said EarthRisk will work with co-founder SEAR Technologies to develop the commercial business model and raise additional capital when the company grows beyond its startup phase.

“Scripps and UCSD have developed some compelling technologies and we are working closely with them to develop business models and to arrange the pieces of a complex puzzle to create successful ongoing commercial enterprises with some of the most promising innovations,” said SEAR Managing Partner JP Plavan.

Another co-founder of EarthRisk is San Diego-based digital-telepathy, a strategic technology partner for software and interface design.

Numerous Benefits

EarthRisk will initially focus on decision makers in energy and commodity investors, but anticipates rapidly developing products for insurers and reinsurers, Bennett said. He said EarthRisk has already secured customer commitments and revenue, but declined to name the first customers.

“A decisions support system like this helps energy companies better manage their resources,” Bennett said. 

Another possible customer base for EarthRisk Technologies is commodities traders, according to Hal DeLong, a business adviser for the von Liebig Center who is coaching Bennett and his colleagues on their startup. 

“People in this business are trying to make a guess on the yields from various crops, weather being an extremely important if not the most important element in the outcome in commercial crop production,” DeLong said.

Bennett added that the financial futures market is also interested in EarthRisk’s technology. “There are groups that trade on the New York and Chicago Mercantile exchanges, for example, who use weather information to forecast what’s going to happen with the use of natural gas,” he said.

Bennett said the next technology step for EarthRisk is to develop its technology to predict severe heat waves in the United States and Europe.

“Phase one of EarthRisk centers on this new and novel research process,” he said. “Phases two and three look at how to make this technology more user friendly and to incorporate multiple data sets. We will focus on developing interfaces that enable decision support using multiple data sets into a system geared toward business users.” 

Bennett envisions an ongoing research relationship with Scripps to keep pioneering new weather and climate research.

Andrea Siedsma is a freelance writer for the San Diego Business Journal.