SAN DIEGO, July 3, 2013 /PRNewswire/ -- EarthRisk Technologies, a pioneer in the research, analysis and visualization of extreme temperatures and weather patterns, released today an important update to its TempRisk solution that features stricter weather input filters and greatly optimized model inputs.
TempRisk v5.0 features increased accuracy and limited complexity by maximizing the quality, not quantity, of model inputs (weather patterns). By largely keeping the number of initial model inputs the same, optimizing the method for determining which of these inputs matter and filtering those that don't, TempRisk v.5.0 better isolates the signal from the noise. Model inputs in v5.0 are based upon these new statistical and meteorological filters applied to both historical and current weather data.
The addition of soil moisture data to TempRisk v5.0 reflects the correlation between soil moisture and air temperature. Dry soils heat faster and more efficiently, which in turn quickly warms the surrounding air. Wet soils heat slower leading to less overall heating over a region. Soil moisture data from 0 to 10 centimeters newly included in TempRisk v5.0 reflects these relationships to produce better temperature models.
TempRisk v5.0 still utilizes a Genetic Algorithm and a Naive Bayes framework that was the foundation of v4.0. A Naïve Bayes framework coupled with EarthRisk's Genetic Algorithm enables TempRisk to predict the probability of extreme temperatures by analyzing multiple weather features that are historically associated with various outcomes.
"We're excited to continue breaking new ground linking atmospheric and data science to provide our clients with uniquely valuable forecast distributions for long-range temperature extremes in the important 2013 summer season," says EarthRisk CEO John Plavan. "With this update we are further refining the accuracy of TempRisk by effectively isolating out more of the input signal and less of the noise that can get in the way. We also added soil moisture data inputs which take into account dry and wet soil conditions and how they directly related to short and long-term temperature patterns."
TempRisk provides the information necessary to extend forecast confidence beyond one week, the point where weather simulations face deteriorating skill. The platform's patent-pending algorithms run millions calculations a day, incorporating thousands of weather patterns, with over 60 years of daily data for multiple global regions. The result: quantified weather risk.
TempRisk is based on the software-as-a-service (SaaS) subscription model and is available now through www.earthrisktech.com.
About EarthRisk Technologies:
EarthRisk Technologies is a San Diego-based software company founded in 2010. The company's product suite, TempRisk, is a web-based platform that crunches more than 60 years of weather data to formulate the odds of cold snaps and heat waves up to 40 days in advance, a lead-time that is twice as long as any in the industry. Trading groups, power producers and energy traders around the globe use TempRisk for a climatological edge in their investments. For more information, visit www.earthrisktech.com
SOURCE EarthRisk Technologies