Task 135

Improved Hypoxia Modeling for Nutrient Control Decisions in the Gulf of Mexico (Hypo-G)

Principal Investigator(s):

J. Stehr


K. Pickering

Last Updated:

October 26, 2012 15:25:36

Description of Problem

Every year, a large dead zone in the Gulf of Mexico emerges, fed by large inputs of nitrogen and other nutrients from the Mississippi river. The dead zone consists of an area of low dissolved oxygen the size of Massachusetts arising from algae blooms and subsequent die off of the algae. Once the algae die, they decay, consuming oxygen and depriving other species of the oxygen they need to live. The dead zone leads directly to fish kills and loss of habitat for less mobile organisms such as mussels and crabs. The goal of the overall project is to provide a scientific basis for addressing the Gulf of Mexico dead zone. In particular, we will use NASA satellite products to evaluate predictions from rain and snow simulations and to provide ocean color evaluations. The University of Maryland (UMD) portion of this project seeks to approach one aspect of the nutrient loading to the Gulf, namely deposition of nutrients from the atmosphere in rain and snowfall. As nitrogen-rich water and snow fall on the watershed of the entire Mississippi River system, a portion of that nitrogen makes its way to the Gulf of Mexico, producing the dead zone there. This project seeks to improve the understanding of these processes and to improve the ability of the models to treat these processes properly. It also responds to a plan put forward by the governors of Gulf of Mexico states, the Governors’ Action Plan for Healthy and Resilient Coasts.

Scientific Objectives and Approach

NASA, EPA and the University of Maryland, along with other partners, have teamed up to address this problem. In particular, our partners are the EPA Gulf of Mexico Program office and the Gulf of Mexico Modeling and Monitoring project at the EPA Office of Research and Development and colleagues at NASA. We are teaming up to use NASA satellite products to help evaluate the outputs of their air and water quality models. The UMD aspect of this collaboration involves comparing NASA and NOAA precipitation satellite and ground-based products to the precipitation part of the Gulf modeling system. This modeling system necessarily includes a meteorological model that drives the air and water quality models. In consultation with leading experts in the field, we have begun using 4 km resolution monthly Parameter-elevation Regressions on Independent Slopes Model (PRISM) data to compare with monthly aggregated WRF precipitation fields. The results are good, but indicate that there is room for improvement. The PRISM model is a regression model that starts from the observational rain gauge data network, and performs an intelligent interpolation between monitoring stations to develop a better representation of precipitation. The dominant parameter is altitude, which is known to strongly affect precipitation, though other parameters are relevant as well.


The first run of the meteorological, emissions and air quality modeling are complete. We have performed some quantitative evaluations of the meteorological model, as mentioned above, and found that the meteorological model performs well in simulating precipitation and somewhat poorer at simulating deposition on a monthly basis. We do not expect perfect performance on a short-term basis, as this requires the model to time precisely precipitation and its intersection with chemistry. We are currently ratioing the precipitation from the model to the PRISM data and will use that to adjust the deposition accordingly. Latysh and Wetherbee (2012), used this approach to improve nitrogen deposition performance.

5 & 6. A conference presentation at the Chesapeake Modeling Symposium on evaluating and improving nitrogen deposition to watersheds is planned for later this spring.

Refereed Journal Publications

Latysh, Natalie E. and Gregory Alan Wetherbee, Environ. Monit. Assess. (2012) 184:913–928, DOI 10.1007/s10661-011-2009-7

Task Figures

Fig. 1 – WRF outputs are shown in a)-c) for a) annual precipitation, b) January precipitation, and c) July precipitation. PRISM data are shown in panels d)-f) for the months corresponding to the WRF results to the left. Areas in WRF with precipitation significantly above that observed in PRISM are highlighted with large + signs; those with precipitation below PRISM observations with – signs. In January, a large area of increased precipitation in the lower Mississippi River Valley is displaced to the north and west of where it appears in the PRISM observations, as indicated by the orange arrow over Alabama and Georgia in the southeast. Overall, such features as the dry wintertime in Florida, the summer monsoon over Arizona, New Mexico and Colorado, and the broad area of summertime precipitation in the Ohio River and Northeast are well represented in the model.