Sheffield Weather

This page shows some recent weather measurements made at approximately 70 metres above mean sea level just to the north of Sheffield in the UK.

Note that, in all of the graphs on this page, the right-most edge corresponds to the time at which this page was downloaded (20:53) and, if everything is working correctly, each graph will show measurements, taken at 15-minute intervals over the past 24 hours, from which the code for this page will attempt to make weather predictions.

Please note that there will be a slight delay while the data are analyzed, plotted (using JPgraph) and then transmitted, so the graphs will appear a few seconds after the text.


The two graphs below show the air temperature as well as the temperature of a horizontal surface exposed to the sky. This surface is black, so will have a temperature similar to a road surface. Both temperature sensors are located about 1 metre above the gutter of a single-storey building. Note that the 18B20 sensor used to measure the surface temperature has failed as a result of water ingress.

When there is a lot of cloud cover, day or night, the air and surface temperatures will be somewhat similar, so the two curves will be closely spaced, if not superimposed. (As the red [surface temperature] points are plotted after the blue [air temperature] points, the blue points may not be visible if the curves are closely similar.) A rise in the surface temperature relative to the air temperature during the day, or a fall in the surface temperature relative to the air temperature at night, should be interpreted as a break in the cloud cover. A sudden and simultaneous dip in both the air and the surface temperatures, especially if the surface temperature dips below the air temperature during the day, is likely to indicate a rain shower. In fact, if the surface temperature is below the air temperature at any time during daylight hours, assume that rain is falling at that time; however, bear in mind that, if the sun is low in the sky, its heating effect on the surface will be minimal. To see an example that provides a better explanation of the potential causes of the temperature difference, click here.

The Maxim DS18B20 temperature sensors used to provide the temperature measurements have a measurement uncertainty of 0.5C (and a resolution of only 0.06C) over the expected temperature range.

Over the past 24 hours, the maximum and minimum air temperatures were 20.9C and 6C, respectively, and the maximum and minimum surface temperatures were 85C and 85C, respectively, as will be seen in the graph below.

The surface temperature (85C) is currently 75.6C higher than the air temperature (9.4C).

As the current temperature of the exposed surface (85C) is not significantly lower than the air temperature (9.4C), the cloud cover must be quite thick.

Atmospheric pressure

The graph below should show atmospheric pressure over the past 24 hours. The pressure sensor is at a height of about 70 metres above mean sea level, but all of the measurements have been corrected to mean sea level for compatibility with other data sources. The Bosch BMP180 pressure sensor used for the pressure measurements has an absolute uncertainty of typically 1mB over the potential measurement range, a relative uncertainty of 0.12mB, and a resolution of 0.01mB. The fine relative uncertainty and resolution mean that the rate of change of pressure can be determined much more accurately than the absolute uncertainty (over the full range of pressure) suggests. If the pressure curve appears "noisy" this probably means that there is a gusty wind.

Standard atmospheric pressure is considered to be 1013.25mB, but the pressure can vary over the range ~950mB to ~1050mB.

If the atmospheric pressure is high, expect fine weather in summer, or cold mainly dry weather in winter; however, if the pressure is low, and particularly if it is falling, expect wet and/or windy weather - especially if the pressure is falling more quickly than "slowly", at which point alarm bells should begin to ring as a storm is very likely to be approaching. The violence of such a storm, and how close it will pass, will depend on the extent, and rate, of the pressure drop.

Wind speed depends on the pressure gradient, but this program knows only the rate of change with time, so can give only a probable wind speed based on the assumption that weather systems are usually moving eastward as a result of the Jet Stream, leading to the pressure varying with time. Therefore, although the predicted wind speed may be approximate, it will provide an indication of the probability and/or ferocity of any imminent storm.

The graph of atmospheric pressure against time, below, may show short/rapid changes of pressure. These can lead to short-duration variations in wind speed (gusts) above, or below, the prediction, which is based on the mean rate of change over a period of 3 hours. These variations in wind speed, lasting for a only a few minutes, can be quite significant, e.g., 25% of the mean wind speed, and are generally caused by turbulence (e.g., caused by the Pennines to the west of this site).

Migraine headaches

Sufferers of weather-induced migraine headaches are unlikely to have a headache in the near future.


The graph below shows the output of a DHT11 humidity sensor allegedly having a measurement uncertainty of 5% over a Relative Humidity range of 20% to 90%. The humidity sensor is in an outbuilding, so will have a larger time constant than with the sensor alone.

The most-recent temperature indicated at the location of the humidity sensor is 18.12C and the relative humidity is reported to be 57%. These suggest a dew point of 9.32.9C. As the air temperature (9.4C) is close to the dew point (9.3C) mist is possible, especially in low-lying areas.


The sensor used to detect the sun is a simple photo-transistor whose output is sampled at one-second intervals by the PIC 16F627 controller that collects the data from all of the sensors. If the brightness is above an arbitrary level (corresponding to little, if any, cloud) this is recorded as such by the controller. Each measurement shown on the sunshine data chart, below, is simply the percentage of the one-second samples, that correspond to bright sunlight, recorded over the previous 15 minutes.