cloudy cloudy

Author Topic: Solar Sensors, Data, and Sunshine Hours - Some Background Information  (Read 740 times)

0 Members and 1 Guest are viewing this topic.

Offline broadstairs

  • Stuart
  • Posts: 6,881
  • Broadstairs, Kent, UK
  • OS/Browser:
  • (Linux Mint)
  • Firefox 51.0
    • Broadstairs
Over the past months or so I have notice that some myths and misunderstandings have arisen over the solar sensor and what is does and I have found myself repeating stuff, so I felt that a new FAQ might be appropriate to cover what a solar sensor is and what it does or does not do.

This FAQ is to try to clear up some of the misconceptions about solar sensors and how they work in relation to Weather Display. I am not covering the set up within WD as Brian has this covered in another excellent FAQ but will try here to dispel some myths and explain how this works.

Firstly people need to understand that the vast majority of solar sensors available for the hobbyist are not accurate instruments. Those from Davis measure global solar radiation on a horizontal plane and not direct radiation. There are other higher cost stations and accurate stations such as Rainwise, Texas Weather Instruments, Campbell Scientific etc.,  although I am not familiar with this equipment. This means that the global radiation sensors cover the whole sky from sunrise horizon to sunset horizon (or as much that is visible from their physical position). Sensors which measure direct solar radiation are in the main very expensive and well beyond the finances of the majority of hobby weather enthusiasts. Basically anyone wishing to obtain reasonably accurate solar readings are left with only Davis or other more expensive stations as choices, the errors created by the cheaper types of sensors are such that this does become rather unreliable.

My experience in all of this is with the Davis Solar Sensor but the principles explained here as to how the software works would apply to any sensor which provides readings in W/mē either directly or converted from some other type of measurement, note that if you are lucky enough to have a sensor which reads direct solar radiation then the rest of this FAQ does not apply.

So having sorted what type of solar sensor the vast majority have lets move on to how Weather Display handles the readings and in particular end up with an estimate of the sunshine hours for a specific location.

To start with there is a maximum solar radiation which arrives at the earth which is about 1366 W/mē measured at the top of the atmosphere, this is then reduced by what is known as atmospheric attenuation or the effect of the atmosphere and a value can be estimated for this effect which reduces the maximum expected value at the earth surface, it has to be an estimate since as we all know atmospheric conditions can and do vary considerably. The actual expected maximum radiation can be calculated by taking account of the latitude and longitude of a location, height above sea level, timezone and time of day which allows calculation of the sun's position (effectively the height of the sun above the horizon represented as an angle), there are also adjustments for the fact that the Earth's orbit round the sun is elliptical and therefore the distance from the sun varies slightly and therefore the maximum possible radiation also varies slightly. This all allows a value to be calculated for the expected maximum solar radiation in W/mē for any location on the planet for any given minute of any given day for a perfectly clear sky with no clouds at all, however remember this is an estimate not an exact value but close enough for our purposes. One other issue with these calculations is that for latitudes above 65° the error in the estimate does increase which can lead to problems for such locations.

Often people will notice that their actual readings will, on occasions, exceed the calculated expected maximum and this is because there is a phenomenon called "edge of cloud effect" which happens when the sun passes through the external edge of a cloud, magnifying the sunlight, this literally behaves like a magnifying glass. It is this effect which increases the readings from this type of sensor potentially by up to 20% , so this is far from unusual and can happen at any time of year.

Next we come on to the area which often causes most questions, the estimation of sunshine hours. As you can see from the previous paragraph there can be no one single value in W/mē which can be used to determine if it is sunny or not. In order to estimate this we use a threshold to detemine if any given minute is sunny or not. This is done by calculating a percentage of the expected maximum W/mē of the current actual W/mē read from our sensor. This works well for most of the day, however the two major downsides of this technique are that for very low sun angle just after sunrise and before sunset both the expected maximum and the actual readings are quite low which gives rise to people often seeing percentage close to or at 100% for about 30 minutes or so after sunrise and before sunset and these very low sun angles are exactly when the solar sensor is least accurate. It must also be said that a lot of people will not have their sensor mounted in a position which affords complete horizon to horizon view which also can impact values when sun angles are low. There is no easy solution to this and explains why Brian has settings on the Solar Control Panel in WD to prevent these low values setting the conditions icons or incrementing the sunshine hours and you should tailor these to your own location with experimentation.

I have found that at least for my location a Sunny Threshold of 85% seems to work best for determining whether or not to increment sunshine hours. I'd suggest a starting point for most locations should be somewhere around 80-85%. Also I have set the times to prevent incrementing sunshine hours or setting the icon to be 15 minutes which is probably a reasonable starting point.


Thanks to Devil, Bashy, Hankster, Brian for their helpful comments  :)
« Last Edit: February 27, 2017, 04:59:14 PM by niko »