First, a warning from my experience: you will not be able to get meaningful data on open source projects' deployments on a large (geographical) scale. What you can do in that regard, though, is to use number of downloads a very approximate proxy measure for assessing the projects' adoption.
Having said that, I think that you can download some information of your interest from those three project repositories and meta-repositories that you have mentioned, but it's not trivial. You have to use a particular set of APIs for each of them. GitHub data is most likely the easiest to access and their set of APIs is relatively simple. OpenHub (formerly Ohloh) left me with an extremely negative impression in terms of their feedback to researchers that wanted to use their "open" data (haven't used their data set). I've been able to use SourceForge data set for my dissertation research, but even that is available to researchers or students under an academic license. For that, you need to request credentials from a specific person at Notre Dame University, which manages the research-focused version of SourceForge data set, called SourceForge Research Data Archive (SRDA).
As a subset of software for my dissertation's data analysis, I have written some open source
R software that works with SRDA (and FLOSSmole meta-repository as well). However, currently the code is not in a form of a standard R package, for various reasons. You are welcome to take a look at it on the following GitHub page: https://github.com/abnova/diss-floss-official (it is not fully and properly documented yet as well). I plan to convert that software into two separate packages - for multi-source open source data collection and for open source ecosystem analysis - if / when I will have a chance. I also recommend you to include CRAN and R-Forge into your data sources, as
R ecosystem contains an impressive number of open source software packages in multitude of scientific domains and I would be very surprised, if hydrology is an exclusion from that tradition.