Models and baby fish – digits, dye, particles, and biological reality

Science Center Objects

July 1, 2016 

By Dr. Robb Jacobson and Ed Bulliner

The Upper Missouri River Pallid Sturgeon Drift Study is ultimately about providing sound scientific information for smart decisions. The decisions relate to how to manage this large river to help recover the endangered pallid sturgeon, or at least to avoid doing additional harm. The Missouri River Recovery Management Plan (http://moriverrecovery.usace.army.mil/mrrp/f?p=136:70) has been developing new sources of science information to address how this can be done.

A prominent hypothesis for the lack of pallid sturgeon population growth in the Upper Missouri River is that there is insufficient drift distance downstream from Fort Peck dam. This idea holds that there is not enough river to allow for development of swimming and foraging capability for the  drifting sturgeon free embryos (the period from hatch until the initiation of feeding) before they are swept into Lake Sakakawea where they may succumb to various causes of mortality, especially the threat of very low dissolved oxygen in the lake.

Dr. Craig Fischenich and colleagues in the US Army Corps of Engineers have developed a computational model – an advection/dispersion model – that can be used to address how river-management decisions change available and required drift distance. It’s a state-of-the-art model, but like any model it has assumptions and estimates. One of the main motivations for the Drift Study is to test the assumptions and find ways to improve the model. The multi-objective structure of the study is meant to determine how well drifting sturgeon free embryos conform to the model assumptions.

Our dye-trace sub-experiment (see previous blog entry Missouri River Dye Trace Experiment to Support Understanding of Free Embryo Drift) is meant to evaluate key components of the advection/dispersion model, in particular how well it predicts transport of a purely passive substance. Early download of fluorometer data(figure 1) from sites within 13 miles of the release site indicate that the dye is moving somewhat faster and is less spread out along the river compared to initial model estimates. Ultimately, passive transport will be compared to free embryo transport to develop biological reality for the model.

Research hydrologist Dr. Susannah Erwin retrieves a fluorometer from the Upper Missouri River to download dye trace data.

Figure 1. Research hydrologist Dr. Susannah Erwin retrieves a fluorometer from the Upper Missouri River to download dye trace data.

(Public domain.)

An important part of the model is the longitudinal dispersion coefficient that measures how particles spread out as they travel down the river. By varying that coefficient we can calibrate the model to more exactly replicate actual conditions. The comparisons between model results in red and dye trace data in blue, before (figure 2) and after calibration (figure 3), show the importance of the dye data. Final calibration of the model will require assessment of a lot more data, but initial calibrations have proven useful in determining where and when fish biologists should sample for pallid sturgeon free embryos during the study.

Comparison between initial model predictions (red) and dye-trace data (blue) about 12 miles downstream of dye injection site.

Figure 2. Comparison between initial model predictions (red) and dye-trace data (blue) about 12 miles downstream of dye injection site.

(Public domain.)

Comparison between calibration model results (red) and dye-trace data (blue) about 12 miles downstream of dye injection site.

Figure 3. Comparison between calibration model results (red) and dye-trace data (blue) about 12 miles downstream of dye injection site.

(Public domain.)