Climate Forecasting

IMPORTANT NOTE: The Climate forecasts I will be making here are not very scientific, as I am not an expert on the matter. Although they are based on real data and (I believe) logic, they are just examples on how you could use KNIME for this kind of research.

This topic elaborates further on my previous Climate Change posts . There you can find some more information on the data sources i have used.

Now,  we will try to make some predictions on the evolution of the maximum temperatures. The workflow below loads the data from ECA_blend_tx directory of files to the Postgres table daily_max_temp. The data contained in these files is the Station Id, the date and the maximum temperature recorded on that date. There is one row for each day.

To know what Station corresponds to each Station ID, I have created the workflow below.  The input comes from the files stations_TNx.txt, also available on the climate website. 

The workflow creates the Stations table in Postgres and contains the station id, station name, the geographical latitude and longitude and altitude of each station

Having this data in place, I will now use the following workflow to calculate the average temperature change over the years.

We will use a subset of our data for a particular station , Getafe. Therefore, I use the following SQL in the DB Query Reader node :

SELECT date, max_temp FROM public.daily_max_temp A, public.stations B where A."station id"

 = B."station id" and B."station name" = 'GETAFE'

Below is a picture of the workflow .

Apparently, the area of Europe around the Mediterranean sea seems to be the most affected by climate change. So now I will try to focus on station data from stations in this area. 

In the workflow below I have attempted to roughly select only the weather stations in Spain which are close to the Mediterranean sea. For the curious, I did this by limiting the Longitude coordinates between - 05 and + 3 degrees.

As you can see in the resulting map below, it is not exact but for demonstration purposes it's close enough.

So now I can select the Stations close to the sea to try to refine my climate forecast.

On this page I  will use some geocoding techniques to more accurately select the weather stations along the Mediterranean coast.

Below is the workflow that does this and below that, the results. The last column indicates that the average yearly temperature increase would be 0,078 degrees. 

(or about 0.8 degrees every 10 years, 4 degrees in 50 years)