The two oil wells located at BP Amoco's Wytch Farm site in Dorset, UK, draw a total of 14,500 barrels of oil a day from beneath Poole Harbour and the English Channel. One well, the M-12, requires an injection of water into the reservoir's lower zones to maintain pressure. But the nature of drilling means that either well could encounter problems which, undetected, could reduce efficiency or lead to costly repairs. Electronic detectors can track well temperature and pressure, but only at the location they are placed. BP and other oil producers are beginning to sample fiber optic sensors, such as those provided by Sensa Ltd., to track temperature and pressure along the entire length of a well. Fiber optic sensors, still an infant technology in this application, provide a complete temperature profile for the oil well in which they are used. For instance, at the Wytch Farm site, they monitor the reservoir's zones in real time to determine which areas are producing and where there might be production problems, such as water breakthrough. Sensa's system uses a surface laser to send a light pulse down the fiber. A computer analyzes the backscattered light from every meter interval of the fiber -- from the surface to well bottom and back again -- and calculates the temperature at 1-m intervals. The system provides accurate temperature to within 0.1 °C. Why is distributed measurement beneficial? "Usually you say, 'OK, I want to look for this feature or that feature,' and you place a sensor at that point," responded Mike Webster, a petrophysicist at BP Amoco's Aberdeen office in the UK. "Data from the length of well is beneficial because you're often limited by your imagination as to what problems there might be. You don't know if there will be a water problem and if you'll spot that. Or it could be an electrical pump. You don't know why a well would become inefficient." Another advantage is the continuous picture that Sensa's equipment provides. By recording a significant temperature drop at the toe of the well, the system was able to track and identify a zone of cold injection water breakthrough, believed to be from a nearby seawater injector. Also, its fast identification of water production from behind the casing of a cemented-off water zone distinguished it from conventional production logging techniques. These methods may have detected the problem, but only after an expensive logging run had been completed and the results interpreted.