ADAPT seminar

Speaker: Dr. Li Li (Department of Civil and Environmental Engineering, Penn State)
Topic: "Data Assimilation for Earth Process Forecasting"
Room: 529 Walker Building (refreshments served)
Time: Monday January 30, 2017 03:00pm to 04:00pm
Abstract: The era of computer-based weather forecasting began about half a century ago. We are now in an exciting time when the earth process forecasting will become possible. Model development have taken giant leaps in integrating earth processes across disciplines. The past decades have also witnessed rapid advances in technology and the generation of novel data at scales as small as nanometers to those as large as the globe (i.e., remote sensing from satellites). In addition, data collection has started to be coordinated through large research community networks including the Critical Zone Observatories (CZOs), the Long-Term Ecological Research (LTER), the Great Lake Ecological Observatory Network (GLEON), the U. S. Geological Survey (USGS), and the National Ecological Observatory Network (NEON). The unprecedented luxury of rich data presents significant opportunities for complex earth process forecasting using models that typically require large numbers of parameters and are challenged by problems including equifinality and parameter uncertainties.

Among a multitude of advantages, the data assimilation techniques can overcome these challenges while at the same time facilitate decisions on observational system design, i.e., what, when, and where to measure. While data assimilation techniques have been widely applied in weather forecasting and hydrology, they have not seen much use in disciplines such as biogeochemistry. In this talk we will discuss opportunities for data assimilation in earth surface process forecasting. We will show the use of the ensemble Kalman filter (EnKF) in assimilating a new type of soil moisture data measured from the intermediate­scale cosmic­ray soil moisture observing system (COSMOS). Compared with point measurements at the centimeter scale, the COSMOS data represent averaged soil moisture at the scale of hundreds of meters. The COSMOS data is assimilated into the land surface hydrology model Flux­PIHM in addition to discharge and land surface temperature observations. We found that the assimilation of COSMOS measurements can improve the model prediction of top layer soil moisture and constrain soil parameters including van Genuchten β and porosity. It however cannot constrain all soil parameters if used alone. The combination of COSMOS and discharge data are important in estimating soil water retention capability related parameters.