if you used the same VNA cal to measure your previous antenna and it was smooth, yet this one is not smooth with the same cal, just switching antennas real-time, then it's not the cal. But it must be something with about 4 nsecs delay as the ripple pattern shows. The antenna is way too small to support that ripple pattern (especially if the simulation does not show such resonances) so other things must be causing something like reflections from nearby (metal desk, metal wall), ground plane effects (if the ground of the coax is not isolated from the new antenna in the same way as the old one), or just a bad spot in the test-cable that comes and goes with flexing (move the cable around with the antenna on it, and see if the ripples come and go). If the ripples come and go with the antenna movement, replace the antenna with the load and move it around again. If the ripples show up in the load it implies a bad cable. If it only shows up with the antenna, then I think it must be some reflection from the environment I don't usually see such ripple with return loss measurements as the range loss is usually big enough to make this effect small; unless your antenna has a lot of gain; but this is wideband so I would expect it to not have much gain. Drop the antenna in a cardboard box lined with absorber and see if they ripples go away (attached to the end of the cable and VNA of course). Where the antenna has a return loss of -19 dB we see a 1 dB ripple, indicative of a secondary signal adding and subtracting with the antenna return-loss at a value of about -37 dBc, which seems quite broadband (the +-0.25 dB ripple near 2.5 GHz at -5 dBc indicates the same level). Is you cal-load maybe not quite 50 ohms (-37 dB ~= 51.5 ohms or 48.5 ohms). But this would only account for the ripple if you calibrated at the end the VNA and then added the test cable after cal.