The midlatitude precipitation seasonal cycle: Gridded Products and Model Assessment

The pronounced seasonality in the midlatitudes offers a unique perspective on the coupling between surface energy and atmospheric moisture budget. Given the fundamental nature of this coupling, we examine the seasonal covariation of surface temperature and precipitation and compare the northern-hemisphere midlatitude zonal mean across observational products, reanalyses, and climate models. We quantify this coupling by their joint evolution in the temperature-precipitation phase space, where they produce a hysteresis loop that serves as a diagnostic and is quantified using the phase lag, lagged correlation, and loop area. We show that not all observational products exhibit the same zonal-mean seasonal cycle, and that the differences among products arise primarily over the ocean. Over the ocean, a four-month lag between temperature and precipitation is a consistent feature in AMIP models, most observational products, and reanalyses. Over land, climate models and products diverge. Products and reanalyses show no lag and a small loop area, whereas climate models show a large spread and a bias in the loop area. A moisture budget analysis comparing climate models and reanalyses suggests that an assimilation term in reanalyses plays an important role over land. Finally, we show that precipitation and evaporation follow a similar seasonal evolution and are highly correlated. Using the bulk formula for evaporation, we find that over oceans, the phase lag arises from a combined contribution of surface winds and water vapor deficit, whereas over land, the main driver is the water vapor deficit. Plain Language Summary Midlatitudes experience a significant seasonal cycle pronounced in both temperature and precipitation, which are coupled by the energy and moisture balances. In this study, we examine the zonal-mean seasonal covariation of surface temperature and precipitation in the northern-hemisphere midlatitudes across different observational products, reanalyses, and climate models. In the zonal mean, the peaks in surface temperature and precipitation have a four-month lag. That said, not all products yield similar results, and satellite-based products exhibit distinct seasonality in the zonal mean. In addition to the observational product spread, we show that climate models exhibit large spread and bias relative to observations over land. Finally, we show that the seasonal coevolution of precipitation and surface temperature can be, to leading order, driven by the seasonality of evaporation. Further examination of the processes controlling the seasonal cycle of evaporation shows different controlling parameters over ocean and land. Over the ocean, seasonality is driven by a combined seasonal cycle of surface winds and water vapor deficit, and over land, the water vapor deficit dominates.