"Prediction time" in the data.gov.uk archive

[damoncl...@gmail.com]

Hi, 

I'm trying to work out how to interpret "prediction time" in the "threehourlyforecast" database at data.gov.uk. I know that is not the same as datapoint but someone at the met office advised me that the data.gov.uk data goes through datapoint and suggested I post here.

Within every "location", "start time" and "time step" combination, there are 6 observations, each with a different value for "prediction time". I'm guessing that these correspond to the hourly forecast iterations within the 6-hour start time block, but I'm not sure how to interpret the actual values of "prediction time". 

To take a random example:

ID	ForecastSiteCode	PredictionId	StartTime	Date	TimeStep	PredictionTime
130130816	3	10acd0b0-65d4-4d0d-994f-19c6fe589b18	0	2016-10-06T00:00:00	0	41
131235028	3	0a9c078e-b2f0-48a2-929e-df33fddab3d5	0	2016-10-06T00:00:00	0	47
130682922	3	af267329-a0eb-4562-9463-d0f60aa6d730	0	2016-10-06T00:00:00	0	63
130406869	3	d92379f5-4193-46e1-b9a9-3fd00a965d59	0	2016-10-06T00:00:00	0	75
131511081	3	79635879-96c7-4c18-9d5e-a4416d933257	0	2016-10-06T00:00:00	0	83
130958975	3	ac4ed6c6-a80b-439a-91dd-2bb2c50c8120	0	2016-10-06T00:00:00	0	95

So what does "41", "47" etc mean? 

Many thanks!

Damon