I take temperature readings using 2 instruments - a Taylor mercury max/min thermometer similair with this one and a Hobo Temp H8 logger. Using the latter, I leave the thermistor wire outside the case for a quicker response. As mentioned on my main data page, these are sheltered from direct and most diffuse solar radiation with air allowed to flow freely around them.

Below are the readings from 2 Hobo Temp loggers placed side-by-side January 28:

They were set to record a data point every 12 seconds from 00:01:23 EST (12:01:23 AM), which is midnight mean solar time at my location). Though the minimum resolution of the H8 is only about .7 °F, it is very accurate and consistent, as can be seen from plots when the day's minimum temperature occurred around 0232 EST (2:32 AM):

Note that they correspond almost point for point, which suggests the actual accuracy is much better than the resolution (because the readings are crossing the threshold between 2 points at very nearly the same time). Logger A is almost new (3rd deployment), and logger B is about 4 years old with a new thermistor wire.

Below are readings around the time of maximum temperature, 1445 EST (2:45 PM):

These vary a bit more than for the minimum, I think primarily because solar radiation does have a small affect (terrestrial radiation does at all times, but is much more consistent) and because of wind gusts. A warm plume or gust probably caused the spikes at 14:44:47 EST, which were 51.79° for logger B & 51.08° for logger A. Any insturment shelter absorbs and emits radiation (though aluminum ones very little) and restrict air flow. So unless the sensor is ventilated, that is probably not much if any better than what I am doing during day and worse during night. The National Weather Service (NWS) does use a ventilated sensor, though to my chagrin I've seen pictures of them from ASOS sites with the ventilating tube pointed more so vertically than horizontally (thus dragging cold air up from the near the surface during a calm, cold night - or vice versa).

The NWS determines daily minimum & maximum temperatures as follows: readings are taken every 10 seconds and quality-controlled with an algorithm to remove data points deemed bad. Those are averaged for each minute (normally 6 readings), and an average of the past 5 1-minute readings is used for the temperature reported any given minute. The extreme values of the temperatures so determined for each minute are reported as the daily minimum & maximum. E.g., the maximum of 48° (F) MPO reported today at 1458 EST (2:58 PM) is the average temperature during the preceding 5 minutes ending at 14:58:00 (2:58:00 PM).

I think the primary reason that is done is for readings in the modern age to be rather consistent with those from earlier times when only liquid-in-glass thermometers were used. I.e., mercury thermometers generally respond to most of an abrupt temperature change in 2-3 minutes, and very nearly all of it in around 5 minutes. If you are outside during a partly sunny day, you'll probably notice the periods which become warmer when the sun peeks in & out of the clouds though - not only the sunshine, but the actual temperature. This often occurs in time periods of much less than 5 minutes. Instantaneous temperatures don't make sense - to the point of absurdity, if the temperature could be measured instantaneously at the molecular level, I am sure there would be some instances of more than 100° most any day. So a time interval must be chosen, but I think 5 minutes is too much for reasons such as stated above. Thus my min/max temperatures are generally an extreme value for about a 75 second interval rather than for 5 minutes ending at a specific minute.

So according to the logged temperatures, for a period of about 92 seconds the maximum was 50.5°. This is, of course, using data without high resolution (NWS sensors have much better resolution and are much more expensive). As best I could estimate, the daily min/max on the Taylor mercury instrument was 30.4°/50.6°. The instruments are not far from piles of snow thrown from the deck, so along with the bit of solar heating, the general spot is probably a bit cool. Thus I decided on a min/max of 30/51 for today - maybe should be 31/51. MPO probably had quite more extensive of a snow cover than my location near the bottom of a local slope facing SE, and they are about 300 feet higher.