SBIR Phase II: Sub-Decadal Weather and Climate Forecast System to Mitigate Risk for Energy and Natural Resource Applications

小型企业创新研究(SBIR)第二阶段项目的更广泛/商业影响是潜在地减少天气和气候对美国能源和保险市场的不利影响。为了帮助美国经济进一步提前为有影响的天气和气候事件做好准备，该项目将开发一个预报系统，可以将气候预报转化为有影响的天气事件的可能性。这些信息将被用来向能源和保险市场通报金融风险，特别是在可再生能源部门，因为天气和气候控制着能源产量。可再生能源是在美国本土生产的，这意味着该项目将通过告知何时何地最容易获得可再生能源来改善国家的能源安全。对该技术潜力的保守估计包括每年为可再生能源行业带来250万美元的收益，对更一般的参数保险市场的影响类似于数百万美元。该项目将增强该公司在NSF SBIR第一阶段奖期间开发的气候预测系统，该系统发布未来十年的气候预测。这些改进包括通过自动机器学习模型参数调整、预报后处理以及将气候预报转换为现实的天气模式来提高预报精度。这项技术使用机器学习来识别陆地、大气和海洋中的模式，这些模式有助于确定气候系统将如何在一个月到十年的时间尺度上演变。然后，这些气候预报将被转化为每日时间尺度上未来天气模式的现实可能性。这些每日天气模式可以用来为许多行业提供可再生能源发电预报和极端天气事件风险的信息，包括能源部门和一般保险业。拟议的预报系统的一个关键技术好处是其高度的计算可伸缩性，这使得能够快速创建通常使用超级计算机产生的气候预报。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

The broader/commercial impact of this Small Business Innovation Research (SBIR) Phase II project is in potentially reducing the detrimental impacts of weather and climate on the United States energy and insurance markets. To help the American economy prepare further in advance for impactful weather and climate events, this project will develop a forecast system that can translate climate forecasts into likelihoods of impactful weather events. This information will be used to inform the energy and insurance markets of financial risk, particularly within the renewable energy sector where weather and climate control the amount of energy produced. Renewable energy is produced locally within the United States, which means this project will improve the nation’s energy security by informing when and where renewable energy will be most available. Conservative estimates of the technology’s potential include a $2.5 million per year benefit to the renewable energy industry, with a similar multi-million dollar impact to the more general parametric insurance market.This project will enhance a climate forecast system developed during the company's NSF SBIR Phase I award that issues climate forecasts up to a decade into the future. The enhancements include increased forecast accuracy through automated machine learning model parameter tuning, forecast post-processing, and the translation of the climate forecasts into realistic weather patterns. The technology uses machine learning to identify patterns in the land, atmosphere, and ocean that help determine how the climate system will evolve on timescales of one month to one decade. These climate forecasts will then be translated into realistic possibilities of future weather patterns on daily timescales. These daily weather patterns can then be used to inform renewable energy power production forecasts and extreme weather event risk for numerous industries, including the energy sector and the general insurance industry. A key technical benefit to the proposed forecast system is its high computational scalability, which enables the rapid creation of climate forecasts that are typically produced using supercomputers.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.