We present the first experimental results demonstrating the use of acoustic tomography to observe the large-scale (>500 km) temperature fluctuations in ocean basins. Use of the tomographic technique, as presented here, will eventually lead to a better understanding of weather, climate ocean circulation, and the distributions of marine organisms. From June to September 1987, tomographic signals were transmitted between four sources and nine receivers over a 3000 by 4000 km region (the basin-scale) in the northeast Pacific. This paper discusses five issues related to transmissions across one 3000 km section. First, basin-scale tomography is feasible because multipaths are stable in the presence of ocean flutations and the multipaths can be understood with ray theory (and its extensions). Second, the tomography problem can be initiated from climatological estimates. It is unnecessary to initiate with conventional measurements during the experiment (a prohibitively expensive proposition). Third, a description of the tomographic reconstruction procedure (Kalman filtering) is presented. The filter imposes temporal and spatial constraints on the solution for the thermal field. Fourth, the sensitivity of the estimated thermal field is examined with respect to variations of the constraints. Thermal maps are not very sensitive to variations of the constraints. Fifth, tomography is used to observe significant changes from month-to-month in the large-scale thermal field in the ocean, including the seasonal thermocline where week-to-week resolution can be obtained. In constrast, available point measurements appear to be insufficient for observing monthly changes in the large-scale thermal field in the ocean. These acoustic measures of weather and climate in the ocean cannot be verified with other data (e.g., satellites, point measurements) and thus provide an orthogonal perspective by which weather and climate may be observed and understood. ¿American Geophysical Union 1991