Preliminary Results of the OASIS Project

The OASIS Project is an enhancement to the Oklahoma Mesonet to allow for routine measurements of net radiation, sensible, and ground heat fluxes at 90 sites across the state. Ten of the 90 OASIS sites, called supersites, also include eddy covariance measurements and thus estimate sensible and latent heat fluxes directly. Installation of all 90 OASIS sites will be completed by 1 January 2000; eight of ten supersites (ALV2, BESS, BURN, FORA, GRA2, IDAB, MARE, and NORM) were operational by 1 June 1999 with the remaining supersites (BOIS and STIG) operational by 1 August. This paper presents results from the ten OASIS supersites during the summer of 1999.

Preliminary analysis of net radiation, sensible, latent, and ground heat fluxes has been examined at 8 supersites for a two-week period. All data were compared using 30-minute means.

Net radiation is measured at each supersite using both an NR-Lite radiometer and a 4-component CNR1. The NR-Lite is a third class system calibrated relative to an Eppley 4-component PSP/PIR system prior to installation. Thus, each OASIS supersite uses two independent net radiation systems. Values of net radiation between the NR-Lite and 4-component CNR1 net radiometers compared well within 5% at all supersites, with most sites within 2.5%.

Sensible heat flux is estimated at the supersites using both profile and eddy covariance techniques. Comparison between the two methods showed generally good agreement under most atmospheric conditions at most sites. Qualitatively, the sonic and profile estimates tracked together well with most correlations > 0.95. During the two week period agreement between the techniques ranged from within 4.4% at BESS to 50% at FORA.

The profile technique estimates the latent heat flux as the residual of the energy budget. Thus, all accumulated error is included in the latent heat flux estimate. Nevertheless, comparisons between the residual and eddy correlation estimates revealed reasonable agreement at most sites. Differences between methods ranged from about 6% at FORA to 42% at ALV2.

Mean daily closure of the energy budget varied between 69% at BESS to 83% at FORA with typical closure of about 81%. Results indicate reasonably good agreement between supersite and standard measurement techniques. Most results varied little from one day to the next, but varied substantially between sites. Thus, while temporal variability was less than expected, spatial variability appears to have greatly affected the accuracy of flux estimates.