The required sampling rate for vertical accel on FDRs is 8 Hz. The actual sampling rate may vary from box to box. For sure, 8Hz is not enough to pick out the true peak 'g' from a hard landing. The load peak from a landing is typically within a quarter of a second from initial contact between wheel and runway surface. But it depends the descent rate... high descent rate landings peak sooner, low descent rate landings peak later. The landing weight and gear type will also make a difference.
In addition, you may also run into filtering problems, as a lot of dedicated FDR accelerometers that I know of have a roll-off around 4 Hz, and it's a rather sharp roll-off. If you have a very high descent rate landing, say above 14 - 16 fps, the frequency content of the load peak may be attenuated significantly.
Using a vertical-speed indicator sampled at only once per second is ill-advised, as the vertical speed will vary a lot in the last seconds before impact, especially in those conditions which are likely to produce hard landings (i.e. steep approach, gusty conditions, etc).
Remember that the may transport category aircraft are certified, what counts is the vertical descent rate at the moment that the tire first makes contact with the runway (of course, certification is in a perfect world, so we assume no bounce after initial contact).