What type of comparisons you would like to do? Comparison itself has no meaning unless you have one of the two groups as reference data. In this case, any software can be used and then you compare the results of the technique or method you utilize keeping in mind that one of them is reference (main data set that you trusted in). If you do not have any trust in the data, you have apply a quality assurance technique before applying any statistical tools.
There are many methods for comparative analysis. The method and software you use depend on what type of comparison you want to make. Some comparison can be done even in Excel. So, depending on your type of comparison you can use SPSS, Systat, OriginPro etc. More over you can try PSPP which is an open source package for statistical analysis. Best Regards!
the most important thing is to compare apples to apples. With respect to atmospheric observations, this means temporal and spatial coverage - if the footprint size of the satellite instrument is 50 km and the ground-based station measures the data in a single spot, the results are not directly comparable. If the satellite provides the measurements for two local times (i.e. 9:30 and 21:10), neither of which overlaps with the local time of dataset B, this is also not a good comparison since the diurnal variation is not taken into account. If one measurement is a result of, let's say, 24h accumulation and the other is an "instantaneous snapshot", this difference also needs to be addressed.
To give a better answer one has to know more about the datasets you want to compare.