9.8 is close enough to 10 for most human scale calculations. No need to have extra sig figs

Interesting that I learned 32.2 ft/s, but only 9.8 m/s - one less significant figure, but only a factor of two in precision (32.2 vs 32 = .6%; 9.81 vs 9.8 is only 0.1%). Here's the fun part - as a practicing engineer for three decades, both in aerospace and in industry, it's exceedingly rare that precision of 0.1% will lead to a better result. Now, people doing physics and high-accuracy detection based on physical parameters really do use that kind of precision and it matters. But for almost every physical object and mechanism in ordinary life, refining to better than 1% is almost always wasted effort.

Being off by 10/9.81x is usually less than the amount that non-modeled conditions will affect the design of a component. Thermal changes, bolt tensions, humidity, temperature, material imperfections, and input variance all conspire to invalidate my careful calculations. Finding the answer to 4 decimal places is nice, but being about to get an answer within 5% or so in your head, quickly, and on site where a solution is needed quickly makes you look like a genius.

Going to guess civil. I work on space systems and we don't have one number. We have the g0 value, which is standard gravity out to some precision, but gravity matters enough we don't even use point mass gravity, we use one of the nonspherical earth gravity models. It matters because orbits.