The following is a status report of the objectives presented in CRiSP's proposal, "Modeling Support Proposal to the Army Corps of Engineers Columbia River Fisheries Programs." This two year contract had been scheduled to begin June 1995 - May 1997, but actually started April 1996.
I. Gas Abatement Modeling
Improved gas generation models were recently developed at the Waterways Experiment Station. These improved models are functions of the volume of spill alone and developed for moderate spills. The gas monitoring and field study data will be used to evaluate the global application of these models. The performance of the new models compared to the old versions of GASPILL will be completed to make precise when these new models should be applied. In conjunction with this work a scheme to choose these models instead of the older versions of GASPILL is being developed and is now in the process of being coded into the next version of CRiSP. CRiSP 1.6 is scheduled to be released sometime in the spring of 1997.
TDG Exposure structure
A new gas distribution function was developed to include lateral mixing as well as the dispersion and dissipation of gas. A number of 2-D mixing and dissipation models were considered. A two flow approach has been chosen in favor of a more complex 2-D model. Prototypes of both were coded and compared. The simpler model has been determined to be satisfactory in summarizing the more detailed model and has been chosen to reduce the computational complexity. This model was presented at the COE Dissolved Gas Modeling Core team in the Portland meeting September 12, 1996.
The simplified 2-D nitrogen model retains the ability to have parcels of spill water travelling through the reservoirs. The appropriate timing of these parcels will be developed from the reservoir specific transect data and velocity estimates from flow data. This calibration work will include comparison of CRiSP velocity data to output from PNL's new river simulation model, see below for details. The 1995-1996 transect and field studies data measuring dissolved gas were obtained and will be used to examine the mixing component of the model. This functionality will also be included in CRiSP 1.6.
With the new TDG exposure equations, the theoretical calculation of fish mortality due to gas needs to be reworked. A prototype using the current model for mortality has been developed, however an improved mortality submodel is also being worked on and is scheduled to go into CRiSP 1.6 along with the new TDG generation and exposure models. The new mortality calculation would be the ultimate result of exposing the fish to the gas dissipation structure described above with an improved mortality model.
In the origional mortality model the mortality occurred as an instantaneous rate of exposure to gas and was calibrated from laboratory experiments. With the improved exposure structure we account for the cumulative effects of gas exposure. The approach is through a threshold model in which mortality occurs once fish have accumulated a lethal number of bubbles. Since fish can accumulate gas bubbles during exposures above some critical saturation level and dissipate bubbles in periods of exposure below a critical level, we have devised algorithms to account for the cumulative time fish spend in water above and below some identified critical gas level. The relationship between bubble accumulation and mortality is being expressed in terms of the vitality based dose-response model (Anderson 1992). The underling mathematical basis for the dose-response aspects of this model has been developed and a publication is being prepared.
II. Hydromodel Improvements
Water temperature changes along different reaches of the Columbia and Snake Rivers were the focus of temperature improvements. Some statistical analysis from the water quality data sets available from the ACOE Gas monitoring reports was completed and compared to previous analysis done on the scroll case data. Forebay temperatures at dams along the Snake and Columbia were compared to those at the upstream reservoir(s). The trends found in temperature were in agreement with the previous analysis. After discussions with the water reservoir control center in North Pacific Division Office of the Army Corps, it was discovered that the source of temperatures in the scroll case data set were variable; the source varied in depth, sometimes in deep water and sometimes in the shallow water of the fish ladder. The warming factors from the more consistent hourly, water quality data set will be used. With this information a "warming" feature will be added to CRiSP's handling of river temperature, where negative warming would represent a cooling effect. Warming factors, related to time of year, will be added to reservoirs where a significant temperature change takes place.
To improve the accuracy of temperature in CRiSP, more upstream data sources were located and the most reliable, extensive data sets will be incorporated into the headwater information in the CRiSP data files. Some of the sources identified were temperature from Anatone on the Snake, the Peck station and Dworshak forebay temperatures on the Clearwater, and Chief Joseph scroll case temperatures. Temperature information from the Snake at several points near Brownlee and Hells Canyon reservoirs has also been attained from IPC.
Both of these changes in temperature modeling will be incorporated into the next version of CRiSP 1.6.
The CRiSP.1 model will be reworked to model the movement and mixing of parcels of water distinguished by temperature and dissolved gas. This function will allow better evaluation of the impact of high gas levels and cooling water releases on specific fish migrations. This new function will be used in conjunction with temperature and dissolved gas modeling improvements.
The water movement algorithm will be calibrated with data developed from Army Corps' Studies. In particular incorporating the work done by Battelle at PNL with regards to their river simulation model will be consulted. In the December 1996 Gas Abatement Model Core Team Meeting, the means and the extent of the incorporation of model information from the river simulation model will need to be discussed. Incorporation of this work could happen at two levels: 1) incorporation of data from actual model simulations from PNL's hydromodel or 2) a recalibration of CRiSP to data produced by these simulations. For instance, we will compare CRiSP's and PNL's simulated velocities to evaluate and improve velocity modeling in CRiSP. This work will depend on the accessibility and quality of simulation data from PNL's model.
III. Internet Support
The following Dissolved Gas Modeling Core Team website has been created to assist in the exchange of information between the modelers: www.cbr.washington.edu/analysis/archive/dgas. Here information about the members of the team and their work is listed and meeting notes and agendas are posted regularly.
As part of the real-time river analysis, data from regional databases will be automatically uploaded into the CRiSP database. This involves the CHROMS and Fish Passage Center databases. We will assist in developing these protocols for communication between the database and the in-season analysis system. In addition, we will assist in coordinating data for calibration of TDG impact models. Gas monitoring and field study data have been obtained from the Corps. Monitoring data has already been made available through DART, but the desire or need for a gas data webpage will be discussed at the December Dgas Model Corps meeting.
The work on gas bubble disease and total dissolved gas (TDG) will require synthesis of the available literature into one database. Dr. Fidler is taking the lead on this effort through a separate contract. We will assist Dr. Fidler and make his data available on our REAL TIME WEB PAGE which is being developed under Bonneville Power Administration (BPA) funding. It contains extensive historical data on river conditions and fish passage. In addition it contains daily information on the status of fish migration.
IV. Analysis Assistance
Our CRiSP staff are ready to assist the Army Corps in the analysis of the impact of river operations on fish survival. For pre- and in-season analysis using projected flow and spill operations we will be able to predict gas levels, water temperatures and survival using CRiSP/Realtime and output from PIT Forecaster. Scheduled for the 1997 season are the incorporation of the improved temperature modeling including reservoir to reservoir temperature changes, gas projections using the new gas generation functions developed at WES, and possibly the new mortality function.
Scheduled to be finished by the end of springof 1997 is the incorporation of the new TDG exposure and mortality models and improvements in water movements to predict the effects of upstream releases from Brownlee or Dworshak. The in-season analysis this year showed a cooling effect linked to the timing of Dworshak releases and so efforts will be made to capture this. Operations data from these projects would be needed to analyze their impact and develop a predictive model.
As part of our support we are developing a PC version of CRiSP1.6. This will be available on the World Wide Web spring 1997. It will be available from the CRiSP model page.