As said: We had a classic British clay brick wall with high Thorium/Uranium content. This produces some discrete lines...
Of course we also have 3 hours backgrounds. We do research not wiki-science...
And no ... sampling for 3 hours instead of 10 minutes does not make discrete lines rise above the continuum background. This is a misapprehension of how randomly arriving signals work. Instead, what will happen is that the discrete lines and the continuum will stay the same size relative to each other but both will become more clearly defined as the bin counts grow (because Poisson noise is less of a factor at high bin occupancy).
It is your lack of understanding of, and control over, noise that invalidates the results you report. If you don't master this now, then it will all come up again when you do new work.