You are here

Home › Forums › W2 Forum › Water Quality

Temperature calibration

Wed, 03/05/2003 - 8:16am — JBlount
I have several questions relating to temperature calibration for anyone that can help me. My attempt at temp calibration of the thermocline in my lake has been fairly successful. However, during times of rapid air temperature increases (e.g., early summer) the modeled temps in the first 4-5 meters of water consistently lag behind the measured temperatures by 1-1.5 degrees. I have eliminated various physical lake parameters as a cause for this lag (e.g., inflow temp and rate). Varying the WSC helps match the shape of the hypolimnion/epilimnion but the temp lag in the first 3-4 meters develops remains after very hot spells. Later in the summer when high temps stabilize for longer periods, the match improves. How can I increase the surface heat exchange system of the model so that the suface water absorbs more heat energy during shorter periods of rapidly increasing air temp (e.g., 2-5 days)? The cloud cover numbers that I am using are the least precise weather data I have. I think that the extent of cloudiness has been overestimated. Would a reduction of solar energy to the water surface due to an overestimation of cloudiness explain this lag in the heating rate even if the air temp was correctly measured and used? Would it be legitimate for me to experiment with reducing the cloud cover factor by 10% to see if it improves the match? Thanks for any help or suggestions. Jon Is it appropriate to vary the wind sheltering coeff at the same water body to reflect seasonal variations in vegetation?[addsig]
Forums: 
Water Quality
Mon, 03/10/2003 - 8:09am — Tom

Jon, You are on the right track. What causes rapid increases in surface temperatures during the day is solar radiation (computed from cloud cover), not air temperature. So, I would fiddle around with the cloud cover values to see what effect this has. Also, wind plays a big role in how rapidly the surface temperatures increase. Less wind allows for a more rapid heating up in both the prototype and in the model. So, you might want to do some sensitivity analyses with these. I also recommend plotting the day before, the day of, and the day after in the model with the observed data. Oftentimes, the temperatures can rapidly change during the time period you are talking about with frontal passages that arrive at the reservoir either before or after they arrive at the met station. Bottom line in all this is that the results are generally clearly a function of the accuracy of the met data. Spending a lot of time getting a better "fit" of the data by adjusting met inputs doesn't give a better calibration than no adjustment. It just points out that the more accurate the input data, the more accurately the model responds. Tom[addsig]