Average surface energy and water budgets, sub-surface variables and atmospheric profiles were computed on-line with an hourly time scale from the ECMWF reanalysis for the five sub-basins of the Mississippi river from 1985-1993. This paper will intercompare the energy and water budgets on diurnal, 5-day, monthly, seasonal and interannual time-scales for these five sub-basins: the Arkansas- Red, the upper Missouri, the upper Mississippi, the Ohio, and the lower Mississippi and Tennessee. The model will be compared with the observed basin scale precipitation and stream flow. The ECMWF reanalysis model shows a realistic seasonal and interannual variability of precipitation, evaporation and soil water. The annual range of soil water is typically 100 mm, and the interannual range is somewhat smaller. The model has a significant spin-up of about 29% in precipitation from the analysis cycle to a 12-24 hr forecast. The spin-up of the model "large- scale" precipitation is 39%, double that of the spin-up of the model "convective" precipitation of 18%. When compared with 5-day and monthly basin averages of hourly raingage observations (corrected for a probable 10% low bias), the precipitation in the reanalysis is low by 20- 25%, while the 12-24 hr forecast precipitation is high by about 5%; so the model precipitation estimates bracket the observations. The nudging of soil water in the analysis cycle, based on 0-6 hr forecast errors in low level humidity, plays an important role in the model liquid hydrology. It prevents the downward interannual drift of soil water, associated with a shortfall of precipitation in the analysis cycle, while allowing realistic interannual variations of soil water. However, the nudging appears to be trying to compensate for other errors in the model; such as errors in the diurnal cycle of low level mixing ratio, and in the seasonal cycle of evaporation. Evaporation in the model is probably high in winter, and on an annual basis may have a very small high bias. The model runoff, which is all deep runoff from the base soil layer, is low by a factor of two, when compared to observed stream flow on an annual basis. The extra Spring runoff is missing. The model diurnal cycle of precipitation has a near-noon maximum, while that observed is late afternoon and evening. This is probably related to the model error in the diurnal cycle of mixing ratio and boundary layer depth. In the model, evaporation and lifting condensation level height are well correlated with soilwater. The interbasin differences in soilwater are realistically linked to precipitation. On interannual timescales the differences in soilwater, evaporation and runoff are smaller than that of precipitation, and seems qualitatively realistic. Overall the ECMWF reanalysis gives a valuable description of the surface energy and water balance of the Mississippi river sub-basins on time-scales longer than the diurnal