Research Interests

I am interested in doing research on the following topics, and they all fall into the general areas of hydrmeteorology, l and surface processes, and land-atmosphere interactions.

• Prediction of hydrologic events at different time scales, from short-term flood prediction to seasonal flood/drought prediction
• Land surface modeling at macroscale suitable for GCM use
• Climate change, land usage change and their relationship

Current Projects

My current research at CEE/AOS/GFDL Princeton University includes seasonal to interannual predictability study and seasonal hydrological forecast, and validation of GFDL land surface model.

Drought monitoring and seasonal hydrologic forecasting

Realtime drought monitoring for the CONUS is carried out weekly using the VIC hydrologic model and the realtime atmospheric forcing from the NLDAS realtime forcing. Seasonal prediction of soil moisture and streamflow is made monthly using our seasonal hydrologic prediction system.

Seasonal to Interannual predictability in the GFDL Climate model

The ability to predict hydrological variables such as precipitation at seasonal to interannual timescale depends on whether there is such a predictability in the climate model. A suite of experiments are being carried out to investigate this question. My collaborators on this project include people from GFDL (Tony Gordon) as well as Prof. Eric Wood.

CEOP Hydrology Reference Sites Survey

Hydrological impact of climate change in the Northeast US

...

Past Projects

In the past few years, I was involved in the following projects and worked on them as part of my Ph.D at Rutgers University.

Evaluation of North American Land Data Assimilation System (NLDAS)

Four land surface models are currectly used in the land data assimilation system. These four models are forced with Eta analysis and observed precipitation and GOES radiation over the North American LDAS domain at 1/8 degree resolution. Are these state-of-the-art land surface model providing similar output? ......

Evaluation of PILPS 2d Land surface models

Validation is one important step in land surface modeling. Comparing model output with proper observations is the major method in validation. Many land surface models were built in the last few decades and different assumptions and simplifications were used. Due to the variety of land surface parameterizations, models behave differently even were forced with exactly the same forcing. PILPS is a project that is aimed to find the reason for the differences and reduce them. PILPS 2(d) experiment, an off-line simulation that forces 21 land surface models with observed meteorology over a seasonal snow cover region is the unique one to intercompare snow models and frozen soil parameterizations...

Land-atmosphere interaction in the Regional Atmosphere Modeling System (RAMS)

The land-atmospheric interaction has become a very hot topic recently. The way to handle land surface in a coupled land-atmospheric modeling system has quite significant impact on the model solutions. A mosaic tiling approach has been accepted in the community to represent the inhomogeneity of the land surface. This approach assumes a homogeneous atmosphere is overlying a heteorogeous land surface represented by different patches, but the atmosphere 'sees' a homogeneous surface since the surface fluxes are spatially averaged over all patches in one grid box. Different from the real world, these two averaging (homogeneous atmosphere and homogeneous surface fluxes) is believed to prevent the mesoscale fluxes. To explore this question, we use a regional modeling system (RAMS) to mimic a mosaic land surface model and overlying atmosphere......

Research Progress Log (personal use)

Write | View