LBA-Hydronet Workshop

Bravo’s recent research includes stochastic rainfall models, rainfall parameterization procedures in time and space, rainfall modeling using a Bayesian approach, disaggregation techniques, downscaling techniques, and large scale climatic influence on local rainfall, with projects both within Venezuela and across the Amazon Basin.

Venezuela’s LBA contribution, SIG-CANAIMA, is the motivation for de Guenni to be spending a year at UNH. She is working with the UNH Water Systems Analysis Group on a project entitled, "Applications of downscaling and disaggregation techniques on studies of the variability of the components of the water cycle in the LBA domain." Currently she is incorporating rainfall data (monthly and daily) into forecast models. The "Red Hidrometeorológica," which supplies the majority of hydroelectric power within Venezuela, has offered some very important data sets. A second source has been a monthly set of accumulated data corresponding to a group of 39 stations in Nebraska, and a third has consisted of output from a dynamic regional climate model in a 4x7 grid over the same area, where grid cells are 40x50 km. Principal objectives of the research are to improve gridded regional climate model predictions, to improve pointwise statistical predictions, and to estimate areal and point rainfall.

Cardoso discussed his work on an Amazonian model that deals with fire’s effects, including changes in albedo, evapotranspiration, and landcover, and consequent changes for hydrometeorology, including cloud formation, rainfall, and runoff. He showed the average precipitation in 1995 from the entire Amazon region and then from areas where the number of fires was >100. The latter were remarkably drier, and in fact in the very wettest areas, virtually no fires occurred.

He plotted fire risk against climate change, and noted that as fire feedback is significant, there are a few possibilities concerning curve shape, which will be the focus of his Ph.D. research. The two variables may have an exponential relationship, but fire risk may tend toward a maximum. Cardoso expressed interest in coupling fire models with ecosystem models, and also in exchanging data with hydrologists in order to augment knowledge in both disciplines.

Chou described the LBA/NH investigation with Yongkang Xue, which uses the CPTEC operational model. A newer version has recently been completed, with slight modifications that she described. She explained the investigation, emphasizing the coupling of the Eta/SsiB model, vegetation map, validations with precipitation observations, and temperature reanalysis. These seem to have the correct temporal pattern for temperature, although peaks are a bit high in the highest temperature locations, but values themselves do not match observations very well. It seems that there should be more cloud cover in the model, and too, there is a very large presence of shortwave radiation that does not appear in the model at all. Because the team wants to run this model inter-annually, it is very important to know the lateral boundary and the physics of the model, because the effects will be amplified.

Cassiano D’Almeida is working with Vörösmarty at UNH’s Water Systems Analysis Group on a water budget closure system for South America and the LBA domain. Key water balance elements are represented by a set of equations focusing on the categories of atmosphere, soil, ground water, and river, all from an aerological approach. They envision LBA-HydroNET as a tool for improved model validation and assurance of data accuracy. D’Almeida gave an expanded presentation on the LBA-HydroNET system the following day during this workshop.

Fekete showed an important organizing database for LBA-HydroNET, the Potential Simulated Topological Networks at 30-minute resolution (STN30p). A higher resolution will be required for application to LBA, targeted now at 6-minute or about 10km. He demonstrated the identification of inter-station regions in the Danube basin, which can be used to compare observed and simulated sub-basin runoff. In some cases the model shows more runoff than precipitation. He showed a comparison among different rainfall data sets for the Tropics, which were very similar in seasonality but which he believes gives estimates that are generally too low. He views runoff and discharge data to be more trustworthy than rainfall data, and the UNH group is developing methods to use runoff to correct precipitation fields in areas of the globe where rivers are uninfluenced by hydraulic engineering (e.g., Amazonia). This work is collaborative with the WMO Global Precipitation Climatology Center and Global Runoff Data Center.

Hurtt presented the Ecosystem Demography (ED) model. ED addresses the specific need for tracking fine-scale heterogeneity in ecosystem structure and fluxes in large-scale studies. The model is designed to represent much of the land-cover heterogeneity that is currently being observed, and retains landscape demography, since forest succession takes time. Causes of heterogeneity may be weather/climate, soils, topography, natural disturbances, or human land use. The core of the ED model is a new scaling method for stochastic point processes. It is essentially a physiologically based "gap model" integrated with other key ecosystem processes including decomposition and nutrient fluxes, as well as disturbance processes such as tree falls, fires, land use and abandonment. The resulting form is a system of partial differential equations for each grid cell. Spatial scales extend from patch to region, and temporal scales from hours to centuries. The model is being developed to utilize data from several scales for model testing or initialization. These data include regional scale datasets, data from state of the art remote sensing sensors, including 1m IKONOS optical data, Lidar structural data on vegetation heights, and GOES fire data, as well as ground based information such as eddy flux, inventories, and experiments from key LBA study sites. ED is also in the process of being coupled to the MM5 meso-scale climate model and a new global climate model for studies of land-atmosphere coupling. Hurtt’s presentation emphasized the need for better coupling between hydrologic studies and ecosystem studies because of the potential for large feedbacks.

Oliveira has been working with Humberto Ribeiro da Rocha at the Universidade de São Paulo (USP) on a project entitled "Turbulent Fluxes of Energy over a Tropical Forest in Amazônia," which constitutes part of the LBA-Ecology project "Measuring the Effects of Logging on the CO₂ and Energy Exchange of Primary Forest in Tapajós National Forest." Objectives of the research are:

• to study the biosphere-atmosphere relationship over a tropical forest area through the balance of momentum, sensible heat, and latent heat fluxes, and
• to collect and analyze the turbulent fluxes and investigate their daily and annual variability.

The experimental site is located in the Tapajós National Forest (TNF), where measurements of wind components, temperature, and water vapor are collected at a height of 65m, as well as soil data: temperature (five levels), moisture (five levels), and efflux. These data are being processed using an eddy covariance system (Moncrieff et al., 1997) in order to estimate mean turbulent fluxes of momentum, sensible heat, and latent heat.

The field experimental data collected will be used to calibrate surface parameters in the Simple-Biosphere model (SiB2) to validate the efficiency of this model over the study area.