The sunniest place on Earth in 2016
What was the sunniest place on Earth in 2016? I used some satellite data to figure it out and here are the results.
Definition of the sunniest place
I defined the sunniest place as the place where you can see the sun for the longest time measured over the whole year (or the largest average sunny hours / day). To see the sun, I required two things:
- The sun must be above horizon
- The sun must not be behind clouds
Data
I used MODIS satellite data from NASA’s Terra and Aqua satellites to get information about the cloud cover. MODIS data was very nice for this analysis as NASA has good MODIS cloud fraction data products publicly and freely available (called MOD06/MYD06). You can check how the data looks like and get more information about the data products, for example, from the awesome NASA Worldview tool.
Terra and Aqua are polar orbiting satellites. This means they fly from pole to pole and scan almost the whole atmosphere every day. The daily overpass times (at all locations, or actually the equator overpass times) of these satellites are
- Terra at about 10:30 AM
- Aqua at about 01:30 PM
As Terra is flying over in the morning and Aqua in the afternoon I used the data in the following way: Terra cloud fraction is valid for whole morning (50% of daily ‘sun is up’ time) and Aqua cloud fraction for whole afternoon (50% of daily ‘sun is up’ time).
To find a simple formula for computing the length of the day (the time from sunrise to sunset given the day of the year and latitude of the location) I just Googled and found a nice paper by Forsythe et al., “A model comparison for daylength as a function of latitude and day of year”, Ecological Modelling, 1995. The paper presented a comparison of some models used for computing the length of the day (or daylength). I just picked the Brock model as it looked very simple and used that. (You can find my Brock model Python code at the end of the document)
Computations
I divided Earth to 1200 x 600 pixels and computed the yearly average sunny hours / day for each pixel. The computations were carried out as follows:
- For each pixel I computed the daylength for each day of the year using the Brock model
- I computed the daily sunny hours value as 0.5 x daylength x Terra cloud fraction + 0.5 x daylength x Aqua cloud fraction
- I computed the average sunny hours over the whole year for each pixel
(To fill the missing data I used the k-nearest neighbors algorithm from the great Scikit-learn machine learning library)
Results
And here are the results:
World
Based on these results, the winner of the sunniest place in 2016 competition is the eastern Sahara desert area (South Egypt and North Sudan)!
Here are the top 10 results:
- Eastern Sahara desert area (South Egypt and North Sudan)
- Yemen and Oman
- Southern Iran and southern Pakistan
- Rest of the Arabian Peninsula
- Namibia
- Northern Chile and southern Peru
- Rest of the Sahara desert
- California (especially the San Francisco bay area)
- Western Australia
- South Africa
I also analysed the sunny hours for different seasons and here are the results:
I thought someone might be interested in closer look of some areas so I ran the analysis with higher resolution for North America and Europe.
North America
Europe
Thanks for reading, hope you liked the images, maybe you can use these results at least for selecting your summer holiday destination. All feedback and comments are very welcome!
And the Python version of my Brock model code is here:
