No abstract

In this paper we present a set of extensions to an existing microbenchmark suite for OpenMP. The new benchmarks are targeted at directives introduced in the OpenMP 2.0 standard, as well as at the handling of thread-private data structures. Results are presented for a Sun HPC 6500 system, with an early access release of an OpenMP 2.0 compliant compiler, and for an SGI Origin 3000 system.

CO 2 and hydrocarbon fluxes from a sitka spruce forest were measured using a conditional sampling method. The method was used in two ways: (1) an automated system was used to monitor continuously the mixing ratio difference of total hydrocarbons and CO 2 in sampling lines for updraught and downdraught air and (2) conditionally sampled updraught and downdraught air was passed through adsorption tubes, which subsequently were analyzed in the laboratory, to determine the fluxes of nonmethane hydrocarbons (NMHCs). Aerodynamic gradient measurements of NMHC fluxes were made over the same period. Method (1) produced a high temporal resolution data set of approximately 4 days of near-continuous fluxes. Marked diurnal trends in CO 2 flux were evident. Peak daylight photosynthetic fluxes ranged from -30 to -90 mmol m -2 h -1. Smaller fluxes were noted during heavy rainfall and diminished photosynthetically active radiation. Nighttime respiration ranged from 0 to 10 mmol m -2 h -1. Total hydrocarbon fluxes were 3 orders of magnitude smaller. During hot, sunny conditions a total hydrocarbon flux of approximately 400 pmol m -2 h -I (CH 4 equivalents) was observed. Fluxes of specific NMHC compounds, from method (2), ranged between 90 and 563 pg m -2 hour -1. The annual carbon flux to Great Britain's conifer forests was estimated to be 1.3 Mt using a simple upscaling model based on the observations of CO 2 flux during the field experiment. Edinburgh, Scotland. 2Institute of Terrestrial Ecology, Penicuik, Midlothian, Scotland. 3Institute of Environmental and Biological Sciences, Lancaster Universit),, Lancaster, England. forests to calibrate CCMs and to estimate the size of the sink are of great importance. Jarvis [1994] described measurements from sitka spruce in NE Scotland but indicated that there were few comparable published measurements of CO 2 excha,e between the coniferous forests and the atmosphere. Biogenic hydrocm'bon emissions also play an important role in atmospheric chemic;try and hence ultimately in the global radiation balance. For example, nonmethane hydrocarbon (NMHC) species represent an important sink for the hydroxyl radical which in mrn controls the atmospheric half-life of radiatively important gases such as CH 4 and oxides of nitrogen [Dw•er e/ al., 19911. However, much uncertainty surrounds the magnitude and distribution of the sources of these emissions on the global scale [Guenther et al., 1995]. Changes in mmospheric temperatures and precipitation patterns may result in enhanced emission of biogenic hydrocarbons w•th potential for complex feedback effects. Predictive modeling of such effects and their influence on global climate requires detailed knowledge of the primary mechanisms which control fluxes. Most of the early work [e.g., Zimmerman. 1979' [Finer et al., 1983] employed branch enclosure techniques wl•ich were scaled to represent the entire canopy. These enclosure techniques are subject to problems associated with heterogeneous canopies and unavoidable microclimatic disturbance within the cha...

The conditional sampling method was used to measure methane (CH4) and carbon dioxide (CO2) fluxes from peatlands in northern Scotland. Preliminary data were obtained using a simple system. Subsequent automation made continuous nux-measurements possible. Observed CH4 fluxes were in the range -70 to +110 mu mol m(-2)h(-1) with a mean flux of 23 mu mol m(-2)h(-1). Peak photosynthetic CO2 fluxes were in the range -10 to -30 mmol m(-2) h(-1). Nocturnal respiration ranged from 0 to +10 mmol m(-2) h(-1). The conditional sampling observations showed reasonable agreement with measurements of flux by eddy-covariance, gradient and aircraft methods. Error analyses and laboratory tests were conducted to determine the precision of the flux-measurement system. The dominant error was associated with the determination of the mixing-ratio difference in conditionally-sampled updraught and downdraught air. The standard error of the difference for CH4 was typically 0.15 ng g(-1) (0.3 parts in 10(9) (p.p.b.))using a careful high-repetition sampling strategy with a modified gas chromatographic/flame ionization detector system. Under typical daytime atmospheric conditions this corresponded to a standard error in the flux measurement of 10 mu mol m(-2)h(-1), which is consistent with field observations. The empirical beta factor in the conditional sampling equation was found to be insensitive to changes in turbulence intensity and atmospheric stability. Simple upscaling models were used to estimate annual carbon-fluxes to Great Britain's peatlands of -0.5 Mt of carbon in the form of CO2 and 18 kt in the form of methane

Measurements of N2O emission fluxes from a 3 ha field of winter wheat were measured using eddy covariance and relaxed eddy accumulation continuously over 10 days during April 1994. The measurements averaged fluxes over approximately 10(5) m(2) of the field, which was fertilised with NH4NO3 at a rate of 43 kg N ha(-1) at the beginning of the measurements. The emission fluxes became detectable after the first heavy rainfall, which occured 4 days after fertiliser application. Emissions of N2O increased rapidly during the day following thr: rain to a maximum of 280 ng N m(-2)s(-1) and declined over the following week. During the period of significant emission fluxes, a clear diurnal cycle in N2O emission was observed, with the daytime maximum coinciding with the soil temperature maximum at 12 cm depth. The temperature dependence of the N2O emission was equivalent to an activation energy for N2O production of 108 kJ mol(-1). The N2O fluxes measured using relaxed eddy accumulation, averaged over 30 to 270 min, were in agreement with those of the eddy covariance system within 60%. The total emission of N2O over the period of continuous measurement(10 days) was equivalent to about 10 kg N2O-N, or 0.77% of the N fertiliser applied