SNTEMP (In)Frequently Asked Questions: Hydraulic Issues

Q44. So how do we know the depth SNTEMP is assuming? Do we have to figure it out arithmetically based on the width and the guesstimated travel time? I only ask this because I can foresee someone asking me "but aren’t all the heating processes affected by depth…?" or some such thing.

A44. In the code, depth of flow is determined by:

In other works, if Manning’s n is less than 1, the programs use Manning’s equation, otherwise, they just back the depth out of the units. And you are right, the solution does depend on depth. I tend to look at it more from the point of view of width because that’s where all the flux is taking place. Others will tend to look at it more from the point of view of depth because that’s where the volume to be heated or cooled is. Thus, what is really important is the ratio of width to depth, because they are both important. [Added 12/2001]

Q121. I am a fish biologist on a national forest in Idaho. I am using your SSTEMP model to predict the effect of water withdrawal on stream temperature. I am doing this by altering the stream inflow and outflow variables. Does the model automatically compensate for changes in width and velocity as inflow and outflow vary?

A121. The answer is a qualified yes. Changes in velocity are "automatically" accounted for since the model "knows" the discharge and the channel's width and length. Velocity is then implicitly available from the equation V = Q/(W*L) where Q is discharge in cubic meters per second and W and L are the channel width and length. Similarly, the Manning's n coefficient also allows calculation of travel time used in calculating maximum daily temperature. Note, however, that in SSTEMP, Manning's n is not a function of discharge, even though it probably should be. If you have actual travel time measurements for different flows, you could develop a much more accurate relationship similar to the width function described below.

Channel width per se is only a function of discharge if you are using a non-zero B coefficient to calculate the width in the relation W = a * Q^B. See the documentation for more about using this function. [Added 6/2002]

Q211. I am preparing to run the SSTEMP model for a stream in the XX Basin. I am at the point of estimating Width's A-Term (I have calculated the Width's B-Term as the slope of the regression of natural log of width versus natural log of flow. In the User's Manual it states that it is best to calculate B and then solve for A in the equation:

W = A * QB

where,
Q is a known discharge, and
W is a known width
B is the power relationship

I am uncertain what values should be substituted for Q and W in this equation to solve for A. I have a series of Q and W data for the site (i.e., the values I used for the regression). I have attached an example of my calculations - would you describe how to calculate the A-term using this equation (rather than using the y-intercept since the manual indicates this relationship breaks down at low flows). Thanks very much for your help,

A211. I believe you have done the right things here. It's been a while since I have done this myself, so I am a bit rusty. The A-term is the EXP function of the regression-derived constant just as you have calculated. I am attaching an old spreadsheet that does basically the same thing you have done. It is, I believe included in the full complement of downloadable SNTEMP files. I suggest that you convince yourself that the B-term applies correctly regardless of the system of units as long as they are consistent, i.e., feet and cubic feet or meters and cubic meters.

Do give some thought to whether you want to use all of your data points in developing your regression. I say this because if it is more important that the model predict correctly at low flows, you may wish to develop the regression based primarily on low-flow widths because you can see from your graph that there is a noticeable departure from a strictly linear relationship on the log-log scale at those low flows. I have no explicit guidance on when you should leave flows out of the regression as that will depend on the flow range that is most important to you. Just document your rationale.

Finally, just as a safety measure, spot check some calculations using SSTEMP to make sure that the estimated widths are appropriate.

Q212. I've been looking over the SSTEMP model for possible use for some simplified TMDLs. In my efforts to become better educated for its possible uses, I have reviewed the available documentation including FAQs. However, I found a description of how water depth is calculated, but I can't seem to find a description of how the flow velocity is calculated when a Manning's n value is used for input. Am I looking in the wrong place?

A212. Good question. Basically I think the answer is reasonably simple. In a unit-less sense, Velocity = discharge/width/depth. Discharge and width are given (or calculated) and depth comes from Manning’s equation.