Really, it is difficult to unambiguously reply to this question. It is known that interpretation of geophysical methods is the ill-posed problem. Each time we need to introduce any available a priori information. And when we have a combination of noise factors …
I propose that selection of method(s) is strongly depends on the concrete physical-geological (physical-environmental) model of the medium under study. In many cases only integration of some geophysical methods from the list: GPR, ERT, magnetic (magnetic gradient), gravity (gravity gradient), self-potential, near-surface seismics, near-surface temperature, etc. may have a necessary effect.
It depends on the target and its depth. In addition to the ground based methods mentioned above, to map the geology and groundwater, airborne methods like liar magnetics, EM and gravity gradiometry can work-- they require corrections for the "cultural noise." You can also use the methods to map buried pipes, dumps, tunnels, pollution, etc.
It is an interesting question.. The nearsurface studies , and not only of subsurface, of the urbanized and metropolitan are very important because the subsurface use is growing for several purposes (civil infrastructure up to tens of meter, resource water and geothermal, characterization and remediation of contaminated sites, structural reconstruction for the geological risk assessment up to several tens of meters). The main problems of the applied geophysics in urbanized areas are the experimental logistic which depend on the operative space, the measurement feasibility which depend on the background noise level (seismic, electromagnetic, temperature,....) and the environmental-human impact of the data acquisition.
I will answer with two experiences in the metropolitan area of Milan Italy where we used seismic method.
The first is regarding the monitoring and control of grouted subway tunnels (Milan metro yellow line) through time-lapse seismic tomography (you can download a paper in my profile https://www.researchgate.net/publication/260576463_Control_of_grouted_tunnels_through_seismic_tomography?ev=prf_pub)
In this case we operated in the night with the stop of traffic with geophone at surface along the street and the source in the tunnel (hole seismic gun in the gallery).
The second experience regards the high resolution seismic reflection using the metropolitan parks. In this case we operated during the day and partially in the night and using a source with low environmental impact (minipulse-an accelerated drop mass). The acquired data are reported in https://www.researchgate.net/publication/258723884_Stratigraphic_evidence_of_a_Middle_Pleistocene_climate-driven_flexural_uplift_in_the_Alps?ev=prf_pub
I propose to the local stakeholder the use of the metropolitan parks (togheter raylways corridors) as a window to explore the subsurface of the metropolitan area.
Finally I think that the innovation in technology and method like land streamer for seismic or geoelectric acquisitions, or low frequency GPR could allow the intensive use geophysics in urbanized area. Generally the urban geophysical acquisition require a dedicated planning and sometime a customization of the instrument and the logistic. We should have to work a lot but it is feasible.
Article Control of grouted tunnels through seismic tomography
Article Stratigraphic evidence of a Middle Pleistocene climate-drive...
Very likely all the mentioned methods can work for a survey of a urban area. Which of them would be the most suitable? And why?
In order to answer to this question we need to know many things, among which: (i) area and max depth of investigation, (ii) portability of the instrumentation, (iii) desired spatial and temporal resolution, (iv) cost we can afford.
The electromagnetic measurements are sensitive to buried elongated conductors such as metallic pipes, electric lines, sanitary sewer, etc. These are usually recognized by the large meter fluctuations which occur in short distance. When the antennas are oriented perpendicular to the elongated conductor’s axis, a minimum conductivity value will be registered (reaching negative values). When the antennas are parallel and close to the conductor’s axis, conductivity values will have a maximum (Geonics Limited, 2010). However, The ERT is an efficient tool for studying the subsoil in urban sites.In general, I propose the use of electric or seismic as more efficient methods that electromagnetic methods. Need to define the study object, study depth and site dimensions.
You may also use ambient vibrations to characterise the subsurface. You measure seismic noise and calculate e. g. H/V ratios (ratio of horizontal to vertical components). See for example:
You can use the IP method with good effect. We achieved very good results with IP inside factory plants. All the pipes and wires will have almost no IP effects. Gravity is also a very good call, together with seismics and ERT. GPR is a very good technique to locate all the "services", which you can use in your interpretation of the other methods to eliminate targets due to infrastructure.
You will definitely have to use more than technique to obtain a good interpretation.
It depends on the target and its depth. the GPR method is good but it requires corrections for the "cultural noise. I am proposing a new method. You can use an electromagnetic device called PL2000 (made in Japen). I've successfully used it ( with Minimum error). that is very better of GPR but only use iron target for example Gas iron pipe , power lines, Conductive cables and ......