The answer to this depends on how many and what type of roofing system you have. Test cuts can be done at different spots on the roof. This is the only certain way to tell if there is moisture in the roofing system(s). Test cuts should be done anyways to determine the make-up of the roofing system(s). Since water runs downhill, test cuts should be taken near drains and gutters, as these areas are the most likely to contain moisture. Test cuts by themselves don’t pinpoint the exact wet areas. For that a moisture scan is needed. We are familiar with two main types – the nuclear scan and infra-red scan. Infra-red scans are done in the early evening as the air temperature cools down. Any wet areas in the roof will retain heat longer than the surrounding areas and an infra-red camera can pick up the thermal differences. A nuclear scan can be done at any time of day. To do the nuclear scan the roof is sectioned off into an imaginary grid, usually 10′ x 10′, and at each intersection point the device is placed on the roof and a reading is taken. The higher the reading the more likely moisture exists. Both moisture scans should be confirmed with test cuts.
Test cuts need to be patched properly to last. Materials similar to the to the roof’s waterproof layer need to be used (asphalt cannot be used on rubber, etc.) and any asphalt patches should use embedded fiberglass mesh. Plastic based membranes (TPO, PVC) should have the seams of the patch heat welded. If the roofing system is under warranty it is usually a requirement to have certain qualified contractors do any work (check the warranty language). The infra-red and nuclear tests have their limitations. The infra-red, which is the least expensive and easiest, does not work well on ballasted single ply systems or show moisture well in anything but the top roofing system if there is more than one. The nuclear is more labor intensive hence costs more. If there are multiple roofing systems then utilizing both provides the most accurate reading.