Chapter 14. Additional Salmon Exercises:
How Should We Measure Habitat Productivity?

Background

On April 14, 1999 the National Marine Fisheries Service (NMFS) issued a biological evaluation of management alternatives for the Federal Columbia River Power System. Their report, titled:

An Assessment of Lower Snake River Hydrosystem Alternatives on Survival and Recovery of Snake River Salmonids

was distributed as one of the appendices of the U.S. Army Corps of Engineers'

Lower Snake River Juvenile Salmonid Migration Feasibility Study.

The Corps study is sometimes called "The Breaching Study" because one of the alternatives is to remove the earth portion of the four dams on the Lower Snake. The dams would be "breached" to allow a faster corridor for the smolts spring migration. The NMFS appendix begins with an interesting collection of time graphs showing the decline in the number of spawners over the past 3-4 decades. The time graphs also show changes in various factors that might be responsible for the decline in spawners. The related factors include the

	number of dams on the river,
	total hatchery releases,
	an index of ocean conditions, and
	a measure of habitat productivity.

The NMFS appendix places much of its emphasis on the first factor. But for these exercises, let's focus our attention on habitat productivity. Specifically, how would we measure the productivity of the Tucannon River? Recall from Chapter 14 that the Tucannon Salmon model represents the habitat by simulating egg deposition, egg loss, emergent fry, juvenile loss, and - at the end of the habitat phase of the life cycle- the number of smolts that begin their spring migration to the ocean. Total productivity might be measured by the number of smolts that begin their spring migration. But what if there were only a few spawners arriving at the spawning grounds? The number of smolts could be quite low, even though the habitat may have the potential to be highly productive.

NMFS deals with this problem by taking the ratio of the number of smolts to the number of spawners. A high value of the ratio is presumably an indicator of a productive habitat. NMFS shows the ratio of smolts per spawner from 1960 to 1993 in Figure 5.3-1. The ratio has varied widely, from low values of around 50 to high values over 150. The average value appears to be around 75.

What can we learn from this data? NMFS used the data to address the possibility that habitat degradation has been a significant factor contributing to the decline in spawners. The NMFS gives its interpretation of the data on page 52:

Although habitat quality is an important factor in salmon demography, the dramatic collapse of spring-summer chinook salmon populations during the mid 1970s is not correlated with reduced smolt-per-spawner ratios. Whereas the annual number of spring-summer chinook salmon returning to spawn declined precipitously in the mid 1970s, there was no concordant precipitous decline in habitat productivity as measured by smolts per spawner.

Discussion
To place the smolt-to-spawner ratio in perspective, consider the ratios from two situations
highlighted in Chapter 14 of Modeling the Environment.

pristine, under populated		Near the top of page 153, we read about 2 thousand spawners at the start of a simulation with pre-development conditions. The simulation begins with a small number of spawners, and the model is used to show "S Shaped Growth" in the population. We might say that the habitat is "underpopulated" at the start of the simulation. Page 153 explains that the 2 thousand spawners lead to 280 thousands smolts that begin the spring migration to the ocean. The smolt-to-spawner ratio is 280/2 or 140.
pristine, fully populated		The bottom of page 155 describes entirely different conditions at the end of the same simulation. (We still have pre-development conditions; there have been no changes in the habitat, in the river corridor or in the ocean). There are around 16.5 thousand spawners which lead to 380 thousand smolts. The smolt-to-spawner ratio is 380/16.5 or 23.

These examples alert us to potential problems using the smolt-to-spawner ratio. In this example the ratio falls from 140 to 23 during a time period with absolutely no changes in the habitat. (The egg loss fraction was fixed at 50%, and the carrying capacity was fixed at 400 thousand smolts.) On the other hand, the "S Shaped Growth" conditions adopted for this simulation may not be representative of the typical conditions expected by NMFS. The following exercises call for you to use the Tucannon model to interpret the likely changes in the smolt-to-spawner ratios that might have occurred over the past several decades.

Five Exercises

1. Add the Smolt-to-Spawner Ratio to the Model:

Figure 14.2 in the book shows the flow of "adults spawning" leaving the system after the eggs are deposited. But if we are to keep track of the smolt-to-spawner ratio, we need to "hold" the value of the spawners until the simulation has had time to hatch the eggs and allow the juveniles to survive their year in the Tucannon. The flow diagram shows how this may be done.

First, add a conveyor stock with a 6 month transit time to "hold" the number of spawners for the 6 month period while the eggs are getting ready to hatch. Then add a second conveyor stock with a 12 month transit time to "hold" the number of spawners for another 12 months while the juveniles in the Tucannon are growing. To get the model started, assume 1 thousand spawners are "on hold" from the past. Set the initial value of the 12 month conveyor to zero . Set the initial value of the 6 month conveyor to 0,0,0,1,0,0.

Then introduce a converter to represent the "smolt to spawner ratio." (To avoid dividing the number of smolts by zero, use the MAX function and divide the number of smolts by 1 or the held value, whichever is larger. Run the model with predevelopment conditions (shown in Table 14.2) and with the harvest fraction at 0. You'll see "spikes" in the ratio whenever the smolts begin their migration. The first few smolt migrations will reveal a smolt-to-spawner ratio slightly above 200. As the population grows "to fill" the habitat, the smolt-to-spawner ratio should fall to around 25 and remain at 25 for the rest of the simulation.

2. How Will the Smolt-to Spawner Ratio Respond to a Degradation in Habitat?
To represent a situation more similar to what NMFS might have in mind, let the model operate with the harvest fraction set to 50% and the smolt migration loss fraction based on 0.90 + RANDOM (-0.025,0.025,123). This will allow 90% loss (as in Table 14.2) in an average year. Losses might be as high as 92.5% or as low as 87.5%. The variation might be attributed to variation in runoff from one year to the next.

Next, set the Tucannon Carrying Capacity as a graph (~) to vary over time. It will be 400 thousand for the first 240 months, decline to 200 thousand over the next 240 months, and remain at 200 thousand for the rest of the simulation. (You might set the length of the simulation to 960 months.) Now, before simulating the model, draw a time graph to show the expected response of the smolt-to-spawner ratio. Will it show a significant decline in response to the degradation in the Tucannon habitat?

3. Simulated Response:
Now, simulate the model and compare the simulated response of the smolt-to-spawner ratio to your sketch. Does the model confirm your expectation? If not, do you need to improve the model or improve your sketch?

4. How Will the Smolt-to-Spawner Ratio Respond when both the Habitat and the River Corridor are Degraded Over Time?

Let's address this question in the same manner as the previous exercise. The harvest fraction will be fixed at 50% and the carrying capacity will be cut in half part way through the simulation. Set the mean value of the smolt migration loss fraction as a graph (~) to vary over time. It will be 0.90 for the first 240 months, increase to 0.94 over the next 240 months, and remain at 0.94 for the rest of the simulation. Now, before simulating the model, draw a time graph to show the expected response of the smolt-to-spawner ratio. Will it show a significant decline in response to the degradation of the Tucannon habitat?

5. Simulated Response:
Now, simulate the model and compare the simulated response of the smolt-to-spawner ratio to your sketch. Does the model confirm your expectation? If not, do you need to improve the model or improve your sketch?