Consulting research done to date will allow you to see how this or similar questions have been addressed by other researchers. Novel ways of addressing the same question are important to move science forward; consulting previous research will help to identify gaps that are opportunities for these novel approaches.
At the same time, consistency in methodology underpins reproducibility, and it's consequently just as important to test the same the questions with the same methods in both similar and novel populations as previous research has done, helping to build a body of evidence and identify if earlier findings are generalizable to other populations.
The problem at hand will come with certain known and unknown elements. In this assignment, depending on the plot of land that you're looking at, you already know certain things about the soil chemistry, geological features or water sources of a given plot.
It is the unknown elements – or a portion of the unknown elements – that your research question will try and address. In this instance, how these factors will impact mung bean production.
A research question that asks
What is the impact of fertilizer x on the growth of mung beans?
would seem a reasonable first attempt at addressing one potential issue at hand. However, it doesn't give a clear definition of what we're measuring as either a dependent or independent variable.
Since the research question informs the hypothesis, which then guides your design, you're leaving yourself with a lot of wriggle room here further down the line.