Summer School on hydrologic assimilation with remotely sensed measurements

July 16-20, 2001
Università degli Studi di Perugia

SHORT COURSE I: Models and Observations for Hydro-meteorological Forecasting (Prof. Dara Entekhabi, darae@mit.edu, 48-331 MIT, Cambridge, Massachusetts 02139 USA)

SHORT COURSE II: Hydrologic Data Assimilation (Prof. Dennis McLaughlin, dennism@mit.edu, 48-209 MIT, Cambridge, Massachusetts 02139 USA)

Data assimilation provides an effective way to combine the models and observations that form the basis for hydro-meteorological forecasting. This summer school consists of two short-courses that consider data assimilation and related modeling and observational issues in an integrated way. Particular emphasis is given to the role of remote sensing and the design of satellite missions. Remote sensing measurements offer a promising new source of information about the land surface. However, this information is usually only indirectly related to variables of hydrologic interest (through nonlinear radiative transfer equations). The problem of estimating hydrologic variables from remote sensing observations often requires the solution of ill-posed inverse problems. Constraints derived from physically-based hydrologic models need to be imposed in order to obtain unique solutions to such problems. Data assimilation provides a theoretical framework and practical methods for addressing such constrained estimation problems.

The first summer school short-course considers data requirements for hydro-meteorological forecasting systems and describes how these requirements can be met with space-borne sensors. The second short-course presents the conceptual basis for data assimilation and describes how ensemble filtering methods can be used to combine hydrologic model predictions with remote sensing data. The two short-courses will share a common hands-on project.

The summer school project considers the problem of mapping large-scale soil moisture from low frequency microwave measurements. Such measurements could eventually be provided by space-borne radiometers such as the forthcoming ESA Soil Moisture and Ocean Salinity (SMOS) mission or the proposed NASA Hydrosphere States (HYDROS) mission. The project will be based on a prototype land data assimilation system which will produce the desired soil moisture estimates by combining synthesized satellite microwave measurements with model predictions and auxiliary data (e.g., soil texture, land cover, precipitation, radiation, air temperature, humidity and winds). Summer school participants will use the data assimilation system to investigate basic design trade-offs and implementation issues. Specific project topics to be considered include 1) tradeoffs between measurement accuracy, frequency, and resolution , 2) adaptive estimation of model errors, 3) sensitivity to incorrect statistical assumptions.

Tentative Program

COURSE I: Models and Observations for Hydrometeorological Forecasting (Dara Entekhabi, MIT) COURSE II: Land Data Assimilation Systems and the Assimilation of Remotely Sensed Measurements (Dennis B. McLaughlin, MIT)

Mon. July 16
2:00-3:15 Lecture I-1: Role of data in numerical weather prediction (NWP); Land-atmosphere coupling and its impact on forecasts; The major land surface processes

3:45-5:00 Lecture II-1: Objectives and methods of data assimilation; Conceptual, computational, and operational issues; State-space formulation of the data assimilation problem.

Tue. July 17
9:30-10:45 Lecture I-2: Modeling and characterization of landsurface hydrologic processes; Sources of error in modeling and error structure; Sources of uncertainty in modeling

11:15-12:30 Lecture II-2: Review of probability and estimation theory, relevance to data assimilation problems.

2:00-3:15 Lecture II-3: Interpolation problems, regression and Bayesian estimation, variational solutions,; applications.

3:45-5:00 Project

Wed. July 18
9:30-10:45 Lecture 1-3: Current approaches initialization of NWP with landsurface states; Remote sensing of landsurface variables; Microwave measurement and radiative transfer

11:15-12:30 Lecture II-4: Filtering problems, sequential estimation; Bayesian methods, classical and ensemble Kalman filtering.

2:00-3:15 Lecture I-4: Algorithms for soil moisture and temperature retrieval based on remote sensing measurements; Sources of errors in retrieval; Systems for making measurements; Introduction of SMOS and HYDROS missions

3:45-5:00 Project

Thurs. July 19
9:30-10:45 Project

11:15-12:30 Project

2:00-3:15 Project

3:45-5:00 Project

Fri. July 20
9:30-10:45 Project presentations

11:15-12:30 Lectures I-5 and II-5: Course and project reviews, open questions and future directions.