I studied pulsars in University and worked with Arecibo, Green Bank, did some data analysis with VLA, and worked on some undergrad projects using low frequency

I studied pulsars in University and worked with Arecibo, Green Bank, did some data analysis with VLA, and worked on some undergrad projects using low frequency arrays. The details get complex, but the gist of it is that arrays let you see bright small scale things, while single aperture telescopes let you see large scale that are dim. Additionally, an array suffers with sensitivity and noise issues if they don't have enough antennas covering discrete distance pairings. This usually occurs at short baselines since array elements often have a minimum spacing. Single aperture telescopes have a large instantaneous gain factor due to their large collecting area, which is really hard to make up without an exceedingly large number of array elements. For example, Arecibo has the same collecting area as ~148 VLA dishes. Data handling for an array is complex as well, as you need to mix and process the data from each telescope to sync them together (correlation), while a single aperture telescope doesn't need this step. On the other hand, an array is generally much less sensitive to instrument noise as the cumulative signal means noise from a single array element has a much lower impact overall. Also, it generally dodges the single-point failure modes by being essentially redundant. You can still operate with one or more elements missing/out of action. Each telescope type has it's pros and cons, and there are equally valid use cases for each one. One thing that I imagine is quite difficult to perform on an array is the type of planetary radar Arecibo specialized in by nature of it's high-power S-Band radar (2.38GHz). With AO's S-Band radar, it was possible to get highly detailed imaging and orbital parameters of near-Earth Asteroids with short observations, while determining orbital parameters with optical observations can take numerous observations (and doesn't generally give you the same detailed 3d models). NASA's Goldstone Observatory is the next best telescope for this, but is nearly 20x less sensitive due to it's size and different operating frequency.