Volcanic eruptions, undersea earthquakes, fissures in the crust, meteoroid accumulation could all be seen as empirical evidence of expansion of the earth. Sink holes and other phenomena could be contrary evidence. Subduction of crustal layers and changes in the earth's rate of rotation could be cited as evidence one way or the other. But the net effect, if measurable at all, might well fall within the range of inaccuracy of the earth's volume estimate. Moreover, nuclear fission produces elements of lesser density. Could this process not produce a reduction of density in the mantle, and hence provide a mechanism for expansion of the earth? I realize that the effect of these factors, although large on a human scale, may well be tiny when compared to the total earth volume. Nonetheless, whatever the measurement or estimate of these factors, how can their effect on the earth's volume, over eons, be considered insignificant, if the accuracy of the total volume of the earth is not known to a degree of certainty?
In short, the current volume of the earth is known to great precision and modern measurements show no significant expansion at meaningful scales.
Facing 20th century evidence that points on the globe -- such as the edges of continents -- that are now distant from each other were once adjacent, some scientists argued that the only explanation was the ongoing expansion of the earth's surface. That argument collapsed by the 1960s, as overwhelming evidence accumulated for the theory of plate tectonics -- that is, roughly, there are relatively rigid continental plates that move across the surface of the earth, with the ocean floor between them created at the mid-ocean ridges and destroyed by subduction under the continental margins. Broad calculations showed that the volumes of the two processes were close enough to cancel each other out.
In the modern era dependent on the accuracy of satellite positions for GPS, sea-level rise measurement, and other detailed geodetic measurements, there was a need to resolve the question of whether the earth's volume was changing in a more conclusive and precise way.
One of the studies that addresses that is the 2011 paper Accuracy of the International Terrestrial Reference Frame origin and Earth expansion, published in Geophysical Research Letters, a journal of the international American Geophysical Union. Most of the paper discusses the technical processes and calculations, but it concludes:
A possible secular change in the mean radius of the solid Earth has long been suspected from various scientific disciplines. The problem was not settled due to lack of a proven theory and the inherent difficulty in actually determining such a change. By using the data combination and the simultaneous global inverse approach, the issue of aliasing from the geophysical processes due to network sparseness is largely overcome here. In fact, our uncertainty in the mean expansion rate determination is dominated by an ITRF [International Terrestrial Reference Frame ] scale drift uncertainty of 0.16 mm yr−1 estimated from comparing VLBI [Very Long Baseline Interferometry] and SLR [satellite laser ranging]. In conclusion, no statistically significant present expansion rate is detected by our study within the current measurement uncertainty of 0.2 mm yr−1.