1

I am trying to create a program that calculates the ideal climate based on a set of inputs and weights. Something similar to 400 cities but based on more climate based options that the user can configure. Some example questions;

  • What is your ideal daily mean temperature, answer must be between 0-40(celsius) ? how important is this, answer 1-5 with 5 being the most important ?
  • How important is it for you to be near the mountains, answer 1-5 with 5 being the most important ?

One of the problems I am having is with the sunshine duration metric as calculated by the Campbell–Stokes recorder.

DISCLAIMER 1: All stats below are approximations / anecdotal DISCLAIMER 2: I may need some statistical calculations explained.

Taking some rough figures from wiki, Los Angeles has 3254 hours of sunshine and London has 1633 hours per year. Maximum sunshine duration in a day is approximately 4380 hours giving LA a sunshine 74% of the days on average and London sunshine 37% of the day on average so we could roughly say that LA has 2x the sunshine hours of London

This does not make sense to me. I want to initially create a metric that more accurately represents the quality of weather. I would then like to refine it to give a metric (configurable) like 'good days' that may need to pass params like > 25C, sunny > 70% of day, cloud coverage < 20%, UV strength > 4.

For example (my brother lives in LA) in August this year there were 25 days in LA that were 'sunny days', lets say sunny for > 70% of the day. In London there were 4, giving values of 83% and 13% approx so we could roughly say that LA has 6X more sunny days per month than London.

My questions are;

  1. Why do the number of sun hours in London seem so high when anecdotally the difference is much much larger.
  2. Where can I find data that is more suitable for my project or alter the data on sunshine hour to suit it.

Many thanks.

Iain Watt
  • 41
  • 2
  • Something to think about. Are the hours of sunshine in the data being confused with hours of daylight? On a cloudy day there's daylight, but no sunshine on the ground. – Fred Nov 24 '17 at 15:32
  • Well the definition of the metric found on wikipedia for the internationally used device is; The Interim Reference Sunshine Recorder (IRSR).[1] In 2003, the sunshine duration was finally defined as the period during which direct solar irradiance exceeds a threshold value of 120 W/m². This does not make sense to me. It seems that the threshold for a sunlight hour is too low somehow. – Iain Watt Nov 24 '17 at 17:32
  • It appears the wikipedia page mentions the metric, but doesn't use it in its calculations. When the sun is below about 4 degrees (rough calculation), the solar energy falls below 120 W/m^2, a bit before sunset. –  Nov 25 '17 at 14:43
  • Note that, in August, the sun in London is up for considerably longer than in Los Angeles, especially in the evenings when people tend to notice. You might need to factor that in. In January, I suspect the reverse would be true (Los Angeles would seem sunnier). –  Nov 26 '17 at 14:35
  • Ok but is there a place to get sunshine hours over a specific value ? so that i can match it with could cover data to get the metric thats useful – Iain Watt Nov 27 '17 at 10:28
  • @IainWatt Thought: combine hourly weather observations (http://opendata.stackexchange.com/questions/10154) with computed sunrise/sunset times (or even "sun-above-x-degrees" times). Still only an approximation, but perhaps a fairly good one. –  Feb 07 '18 at 13:32

0 Answers0