Evaluations of CALPUFF, HPAC, and VLSTRACK with Two Mesoscale Field Datasets

Abstract

Results of evaluations of transport and dispersion models with field data are summarized. The California Puff (CALPUFF), Hazard Prediction and Assessment Capability (HPAC), and Chemical/Biological Agent Vapor, Liquid, and Solid Tracking (VLSTRACK) models were compared using two recent mesoscale field datasetsthe Dipole Pride 26 (DP26) and the Overland Along-wind Dispersion (OLAD). Both field experiments involved instantaneous releases of sulfur hexafluoride tracer gas in a mesoscale region with desert basins and mountains. DP26 involved point sources, and OLAD involved line sources. Networks of surface wind observations and special radiosonde and pilot balloon soundings were available, and tracer concentrations were observed along lines of whole-air samplers and some fast-response instruments at distances up to 20 km. The models were evaluated using the maximum 3-h dosage (concentration integrated over time) along a sampling line. It was found that the solutions were highly dependent upon the diagnostic wind field model used to interpolate the spatially variable observed wind fields. At the DP26 site, CALPUFF and HPAC had better performance than VLSTRACK. Overall, the three models had mean biases within 35% and random scatters of about a factor of 3-4. About 50%-60% of CALPUFF and HPAC predictions and about 40% of VLSTRACK predictions were within a factor of 2 of observations. At the OLAD site, all three models underpredicted by a factor of 2-3, on average, with random scatters of a factor of 3-7. About 50% of HPAC predictions and only 25%-30% of CALPUFF and VLSTRACK predictions were within a factor of 2 of observations.

A Review of Methodology for Evaluating the Performance of Atmospheric Transport and Dispersion Models and Suggested Protocol for Providing More Informative Results

“…These typically include, the normalized mean square error (NMSE), the fraction of predictions within a factor of 2 of the observations (FAC2), the fractional bias (FB), the geometric mean bias (MG) and the geometric variance (VG) (see Appendix A for definitions). In the absence of any universally agreed performance criteria, when authors wish to compare their results to those of others, they generally cite the criteria for an 'acceptable model' proposed by Chang and Hanna (e.g., [16][17][18][19][20][21][22]), which are summarized in Table 2. The Chang and Hanna criteria were based on their experience in conducting a large number of model evaluation exercises [23].…”

Nuclear tools for characterising radiological dispersion in complex terrain: evaluation of regulatory and emergency response models

“…The FA2 values for LH range from 57-71% depending on the stability, which is very high compared with results from Terada et al (2004) who reported only 33% for the new version of WSPEEDI (a regional model simulation). Our models also compare well against the three models evaluated by Chang and Franzese (2003), who found FA2 results of 52%, 60% and 43% for CALPUFF, HPAC and VLSTRACK models respectively. However, both of these studies were conducted on much larger scales, so the current results are unsurprising.…”

“…Recent dispersion model evaluations using nonradiological species include the study of Chang and Franzese (2003) which compares the California Puff (CALPUFF) model, the Hazard Prediction and Assessment Capability (HPAC) and the Chemical/Biological Agent Vapor, Liquid, and Solid Tracking (VLSTRACK) model using data from a recent mesoscale field campaign (Dipole Pride 26, DP26) in which 30 air samplers measured 15 min-average concentrations of SF 6 over a three-hour period. However, only hourly averaged concentrations were used in their results, since CALPUFF cannot produce higher frequency data.…”

Development of NEXRAD Wind Retrievals as Input to Atmospheric Dispersion Models

“…To assess whether the conclusions regarding the effect of the NEXRAD wind retrievals on the overall dispersion presented in Section 5.3 were specific to CALPUFF, we also performed a series of simulations using the Second-order Closure Integrated PUFF (SCIPUFF) model (Sykes et al 1993(Sykes et al , 2004). The SCIPUFF model performs the dispersion predictions as part of the Defense Threat Reduction Agency's (DTRA's) Hazard Prediction and Assessment Capability (HPAC) used by the U.S. Department of Defense (DTRA 2001;Warner et al 2004;Chang et al 2003Chang et al , 2005. CALMET and SCIPUFF are also being used to represent dispersion for the Biological Warning and Incident Characterization (BWIC) system currently being developed by DHS.…”