Atmospheric moisture residence times and cycling:

Implications for how precipitation may change as climate changes

Kevin E. Trenberth

National Center for Atmospheric Research1
P. O. Box 3000
Boulder, CO 80307


fax: (303) 497 1333
ph: (303) 497 1318

To appear in forthcoming August 1997 GEWEX (Global Energy and Water Cycle Experiment) News

1 The National Center for Atmospheric Research is sponsored by the National Science Foundation.

Characterizing all aspects of the hydrological cycle accurately from observations and analyses is a difficult task, so that there remain substantial uncertainties in precipitation, evaporation, the moisture transport in the atmosphere and surface runoff. These uncertainties become magnified in attempts to project what changes may occur in any of these quantities as the climate changes. Nevertheless, the availability of new datasets from GEWEX and from the reanalyses have allowed new estimates to be made of quantities important for understanding the hydrological cycle and its role in climate change. In this study (Trenberth 1997) we make use of the monthly Climate Prediction Center (CPC) Merged Analysis of Precipitation (CMAP) fields of precipitation P from Xie and Arkin (1996, 1997), precipitable water from NVAP (Randel et al. 1996) based primarily on Special Sensor Microwave Imager (SSM/I) over the oceans and rawinsonde measurements over land plus TIROS Operational Vertical Sounder (TOVS), and evaporation E and moisture transport from the National Centers for Environmental Prediction (NCEP)/NCAR reanalyses.

In particular, we have made new estimates of the e-folding time constants of moistening of the atmosphere through evaporation at the surface and of the drying through precipitation. For precipitation, local values of the depletion rate of atmospheric moisture are about 1 week in the tropical convergence zones but they exceed a month in the dry zones in the subtropics and desert areas. The restoration rate shows largest values over northern Africa, Saudi Arabia, Iran and Australia. Values average about 12 days in the tropical convergence zones and are lowest in the subtropical highs where evaporation is a maximum, yet moisture is trapped at low levels by subsidence and so precipitable water is limited. Overall globally, the e-folding residence time for atmospheric moisture is just over 8 days. Note that this is the global mean of the inverse time constant expressed as a residence time, not the residence time computed from global means (which is a day or so longer).

These computations do not take into account atmospheric moisture transport. Therefore we have also made new estimates of how much moisture that precipitates out comes from horizontal transport versus local evaporation, referred to as ``recycling''(see Brubaker et al. 1993, Eltahir and Bras 1996). The results depend greatly on the scale of the domain under consideration. A new formulation of the recycling allows it to be computed locally and mapped by making assumptions of uniformity (that do not exist in practice); however, the very heterogeneity of the land surface and the catchment basins makes this approach useful. We have produced global maps of the recycling for annual means for 500 km and 1000 km scales (Fig. 1, see below for PostScript or gif version ). For the 500 km scale, global annual mean recycling is 9.6%, consisting of 8.9% over land and 9.9% over the oceans. Over the Amazon, the average is about 5% and over the Mississippi basin about 7%. For 1000 km scales the mean recycling is 16.8% globally, 15.4% over land and 17.3% over the oceans. Over particular river basins, our results are not incompatible with those of previous studies. It is worth pointing out that the larger values previously obtained for the Amazon versus the Mississippi are mostly a result of the scale of the domain, (see Brubaker et al. 1993). Thus we find that even for 1000 km scales, less than 20% of the annual precipitation typically comes from evaporation within that domain.

While overall atmospheric moisture depletion and restoration must balance, precipitation falls only a small fraction of the time. Thus it is important to also consider the depletion rate conditional on when it is raining and so we also examine precipitation rates (also called precipitation intensity). Over the United States we use a dataset from Higgins et al. (1996). One hour intervals with 0.1 mm or more are used to show that the frequency of precipitation ranges from over 30% in the Northwest, to about 20% in the Southeast and less than 4% just east of the continental divide in winter, and from less than 2% in California to over 20% in the Southeast in summer. Overall, in midlatitudes, precipitation typically falls only 10% of the time, and so rainfall rates, conditional on when rain is falling, are much larger than evaporation rates.

All the above highlight the mismatches in the rates of rainfall versus evaporation which imply that precipitating systems of all kinds feed mostly on the moisture already in the atmosphere. Over North America, much of the precipitation originates from moisture advected from the Gulf of Mexico and subtropical Atlantic or Pacific a day or so earlier.

How should rainfall, or precipitation, change as climate changes? Why are the patterns predicted from different models under increased greenhouse gas scenarios so different? What is the relationship among changes in evaporation, changes in moisture content of the atmosphere, and changes in precipitation? What are the factors that should be taken into account to explain the changes? The IPCC 1995 report (IPCC 1996) in dealing with future climate prospects with increased greenhouse gases in the atmosphere states that ``Warmer temperatures will lead to a more vigorous hydrological cycle; this translates into prospects for more severe droughts and/or floods in some places...'' ``Several models indicate an increase in precipitation intensity, suggesting a possibility for more extreme rainfall events.'' Therefore, we have made use of the above results in an attempt to explain the processes involved and address some of the questions. In particular, we note the importance of examining rainfall rates (or intensity) and rainfall frequency, not just accumulated amounts.

Increases in greenhouse gases in the atmosphere produce global warming through an increase in downwelling infrared radiation, and thus not only increase surface temperatures but also enhance the hydrological cycle, as much of the heating at the surface goes into evaporating surface moisture. Global temperature increases signify that the water-holding capacity of the atmosphere increases and, together with enhanced evaporation, this means that the actual atmospheric moisture should increase. It follows that naturally-occurring droughts are likely to be exacerbated by enhanced potential evapotranspiration. Further, globally there must be an increase in precipitation to balance the enhanced evaporation. Observations confirm that atmospheric moisture is increasing in many places, for example at a rate of about 5% per decade over the United States (Ross and Elliot 1996). Based on the above results which show that perhaps 70% of the moisture in an extratropical storm comes from moisture stored in the atmosphere, we argue that increased moisture content of the atmosphere therefore favors stronger moisture convergence in all precipitating weather systems, whether they be thunderstorms or extratropical rain or snow storms. This leads to the expectation of enhanced rainfall or snowfall events, thus increasing risk of flooding, which is a pattern observed to be happening in many parts of the world (Karl et al. 1996).

Moreover, because there is a disparity between the expected rates of increase of atmospheric moisture and precipitation (the latter is limited by the surface heat budget; see Trenberth 1997 for details), there are implied changes in the frequency of precipitation and/or efficiency of precipitation (related to how much moisture is left behind in a storm). In particular, if rain rates increase faster than rain amounts, then the frequency of rain could decrease. However, an analysis of linear trends in the frequency of precipitation events for the United States corresponding to thresholds of 0.1 and 1 mm/h shows that the most notable statistically significant trends are for increases in the southern United States in winter and decreases in the Pacific Northwest from November through January, which may be related to changes in atmospheric circulation and storm tracks associated with El Niño-Southern Oscillation trends (Trenberth and Hoar 1996).

The above arguments suggest that there is not such a clear expectation on how total precipitation amounts should change, except as an overall average. Therefore, we suggest that a great deal more attention should be paid to the rates (or intensity) of precipitation, both in observations, and in models, conditional on when it is falling, and the frequency of precipitation. It is further suggested that the focus should be on 1-hour average rates as a useful compromise that is reasonably compatible with the lifetime of the main precipitating systems in nature, but which goes beyond instantaneous values, and is feasible from global climate models which typically have time steps of about half an hour. This would facilitate a detailed analysis of the diurnal cycle of precipitation both in models and in nature which should be very enlightening. These topics should fit well under the GVaP program.


Brubaker, K. L., Entehabi, D., and Eagleson, P. S., 1993: Estimation of continental precipitation recycling. J. Clim., 6, 1077-1089.

Eltahir, E. A. B., and Bras, R. L., 1996: Precipitation recycling. Rev. Geophys., 34, 367-378.

Higgins, R. W., Janowiak, J. E., and Yao, Y-P., 1996: A gridded hourly precipitation data base for the United States (1963-1993). NCEP/Climate Prediction Center Atlas No. 1. U.S. Dept. Commerce. 47 pp.

IPCC (Intergovernmental Panel of Climate Change), 1996: Climate Change 1995: The Science of Climate Change. Eds. J. T. Houghton, F. G. Meira Filho, B. A. Callander, N. Harris, A. Kattenberg, and K. Maskell, Cambridge Univ. Press, Cambridge, U.K. 572 pp.

Karl., T. R., Knight, R. W., Easterling, D. R., and Quayle, R. G., 1996; Indices of climate change for the United States. Bull. Amer. Meteor. Soc. , 77, 279-292.

Randel D. L., Vonder Haar, T. H., Ringerud, M. A., Reinke, D. L., Stephens, G. L., Greenwald, T. J., and Combs, C. L., 1996: A new global water vapor dataset. Bull. Amer. Meteor. Soc., 77, 1233-1246.

Trenberth, K. E., 1997: Atmospheric moisture residence times and cycling: Implications for rainfall rates with climate change. Climatic Change (submitted).

Trenberth, K. E., and Hoar, T. J., 1996: The 1990-1995 El Niño-Southern Oscillation Event: Longest on record. Geophys. Res. Lttrs., 23, 57-60.

Xie, P, and Arkin, P. A., 1996: Analyses of global monthly precipitation using gauge observations, satellite estimates, and numerical model predictions. J. Clim., 9, 840-858.

Xie, P, and Arkin, P. A., 1997: Global precipitation: A 17 year monthly analysis based on gauge observations, satellite estimates and numerical model outputs. Bull. Amer. Meteor. Soc. (submitted).

Fig. 1. The recycling in percent, for annual mean conditions, computed for a length scale of L = 1000 km, and using E and moisture fluxes from the NCEP reanalyses and P from CMAP.

PostScript Version (1.1 MB) Typically will be loaded by ghostview or other default PostScript viewer on your system. Alternatively, if you are using Netscape one may press the shift key and then simultaneously click on the ``PostScript Version'' link to download the PostScript file.

gif Version (15.1 kB) You may use the "File", then "Save As ..." option of the Netscape browser to download this image.