I made ambient temperature and relative humidity measurements using a sensor, next to a highly accurate meteorological station and i want to test the accuracy of my data in respect to the meteorological station (assumed as the "true" value).

Practically, I have 2 sets of data, one is the observed, which is consisted on one measurement per 10 minutes for about 20+ hours (total of 130~ measurements, one for every 10 minutes), and similar data from the meteo station.

I am not sure regarding the statistical analysis of the data, since every measurement has different value because it was taken in different time (10 min interval).

Should the analysis be made on the residuals?
Should I use the absoloute value of the residuas in the analysis?
Do I need to use RMS? or STD?
Is it important to define the distribtion of the data? (data = residuals? or measured values)?
Can you recommned on similar experiment to look for referenceq/guidance?

The objective is to analze the accuracy/precision/reliability of my measurements in respect to the highly accurate meteo station data.

thanks advace.