Re: Does radiation slow down ice forming under bridges?

You are probably correct that the downwelling longwave (infrared) radiation from the underside of a bridge is greater than the downwelling longwave radiation from the sky, and that this leads to a warming effect on the water under the bridge. However, there are several factors which will tend to offset this effect; we must compare their magnitudes.

Let's try to find some numerical values for the longwave reduction. I'll take a winter surface temperature of -5 C, which is a decent average daily winter temperature here in Boston. Suppose the bridge covers most of the sky visible from the water under it. Suppose also that the bridge acts as a blackbody. (a polished metal surface is far from a blackbody, but bridges are painted or rusty.) The downwelling IR radiation flux (in Watts/m^2) will be given by

LW = sigma T^4

Under clear-sky conditions, you might expect zero longwave radiation to be coming down from the sky. This is not the case: emission by greenhouse gases in the atmosphere is very roughly equivalent to the emission by a blackbody at a temperature of 255 K on the global average -- perhaps closer to 245 K in the wintertime. This gives a clear-sky downwelling longwave of LW(clear) = 204 W/m^2. The underside of clouds at 2 km altitude might be 255 K in the winter, giving LW(cloudy)=240 W/m^2. So the effect you mention amounts to a maximum change of delta-LW = 90 W/m^2 in clear-sky conditions, or 50 W/m^2 in cloudy conditions.

But the bridge will block out the incoming sunlight striking the water! This reduction in solar heating will tend to cool the water. At 45 north latitude in January, the daily-average solar incoming radiation is 140 W/m^2. Water reflects only 10% of this 125 W/m is absorbed. If the bridge really does entirely shield the water from the sky, reduction in incoming solar will entirely cancel the increase in incoming longwave radiation. If the bridge covers half the sky, both factors will be reduced by very roughly a factor of two (this depends on the orientation of the bridge with respect to the sun.) If it's cloudy, the reduction in incoming solar between the area under the bridge and the area away from the bridge may be as low as 20 W/m^2, but this is still enough to reduce the difference in longwave of 50 W/m^2.

You also mention the convection-limiting and wind-sheltering effects of bridges. I think the former effect will only be important on perfectly calm days. Normally, the wind blows at a few m/s, while the rising motion of convective parcels is a few cm/s near the ground. Unless the bridge is much, much wider than it is high, by the time a parcel of air has risen from the water to the height of the bridge, it will have been blown away from the bridge by the wind.

As for the wind-sheltering effect, in my experience, it's actually windier under a bridge than elsewhere. This is because air striking the bridge is forced to flow over and under it. I often see ripples and wind-streaks under a bridge when it's calm elsewhere.

I believe this is the key to the puzzle! I've observed that fresh water in windy conditions tends not to freeze, for two reasons. First, the constant motion of the ripples on the surface tends to stop small ice crystals from forming, and to keep them from joining once they've formed. Second, below 4 degrees C, colder water is less dense than warm water. So, in calm conditions, the surface skin cools rapidly, but does not mix with the warmer water below, because it's less dense. But when it's windy, the turbulence caused by the wind mixes the cold water with the warm water beneath it, which stops the surface from becoming cold enough to freeze.

You should be able to do experiments to test these ideas. To test your idea (that changes in longwave radiation reduce freezing under bridges), take two large buckets of water and insulate their sides. Support a large metal sheet (a garbage can lid, perhaps?) a meter or so above one bucket. (If the sheet is too close to the bucket, it will affect the wind patterns.) See if one freezes before the other. To test my idea (that the wind-generated ripples inhibit freezing), take the same two buckets without the metal plate, but make ripples in one bucket. You can do this by making a little stirring device attached to an electric motor, or you could use an aquarium air pump to make bubbles in the bucket.

References:
Peixoto and Oort, Physics of Climate, American Institute of Physics, 1992
Atmosphere-Ocean Dynamics, Academic Press, 1982

---