The Passive Microwave Water Cycle Product published by Remote Sensing System in July 24, 2009.
The purpose of this technical report is to document the methodology we use for producing
the Passive Microwave Water Cycle (PMWC) product. The PMWC product provides a full
characterization of the atmospheric branch of the water cycle over the ocean, and it includes the parameters: water vapor, water vapor transport (speed and direction), water vapor transport divergence, evaporation, and precipitation. This product was developed for the NASA Energy and Water Cycle Study (NEWS) and is an extension of the work by Wentz et al. (2007).
This report specifically describes the methodology used in producing the Version-1b PMWC product. The product has a spatial resolution of 0.25-degrees, a temporal resolution of one month, a time period from July 1987 through December 2006, and coverage over the global oceans observed by the SSM/I sensors (F08, F10, F11, F13, F14, and F15).
The PMWC product is available in a simple binary format. A “readme” text file describing
the format is provided. We also provide FORTRAN, IDL, and MatLab read code and a text file to verify that the read code is working properly. The product is available on Remote Sensing Systems’ (REMSS) public FTP server (ftp://ftp.remss.com/water_cycle/), a link to which can be found on the REMSS homepage (http://www.remss.com).
The evaporation calculations require an estimate of sea surface temperature (SST). Retrieval of SST for water warmer than 12°C is only possible using X-band channels (10.7 GHz). For colder water, C-band channels (6.9 GHz) are required. Thus, tropical SSTs are available starting in 1998 with TRMM and global SSTs are available starting in 2002 with AMSR-E. Since neither of these products is available before 1998, and in order to have consistent SST input, we use instead the NOAA/NCEP Reynolds optimal interpolation (OI) Version-2 SST (Reynolds et al. 2002). While of limited resolution, it is a widely used and quality-checked product available throughout the time period needed for construction of the PMWC dataset.
Evaporation is primarily driven by wind speed and SST, but the calculation also requires
estimates of relative humidity and air-sea temperature difference. We obtain relative humidity from the International Comprehensive Ocean-Atmosphere Data Set (ICOADS) and air-sea temperature difference from the difference between the Hadley Center’s Marine Air
Temperatures (MAT) and Reynolds SST. Only nighttime MATs are used to avoid biases due to solar heating of ship decks. For both datasets, we average over the 19 years 1987-2006 and make 12 monthly climatological maps. The advantage of using a climatology is to prevent introducing spurious trends in relative humidity or air-sea temperature difference to our product that are hard to validate. We note that ICOADS relative humidity has no appreciable trends above the noise level, while the Hadley Center’s MATs have a well known cooling trend (Christy et al. 2001). Full details are available in Wentz et al. (2007, SOM). Since Wentz et al (2007) were concerned with global average values and not spatial patterns, they did not spatially filter these climatological maps in any way. Since the PMWC product is at a higher resolution, and since spatial patterns are important, we applied the additional step of smoothing these climatological maps by fitting with spherical harmonics. Examination of spatial patterns of evaporation suggests that use of climatological information about relative humidity and air-sea temperature difference underestimates the atmospheric variability in meteorologically active regions (Pete Robertson, unpublished work).