Air-surface exchange of mercury (Hg) above an arid geothermal area was measured with three parallel flux chamber experiments. The different experimental designs were intercompared with each other, with regard to the magnitude of the measured Hg fluxes and their response to environmental changes. Qualitatively, the measured Hg fluxes agreed well throughout the diurnal cycle, and in their response to environmental events and experimental manipulations, but quantitatively, there were significant discrepancies between the individual flux results. On average, the three designs yielded Hg fluxes agreeing within a factor of 2, but even more pronounced differences were observed during midday high emission periods and during apparent nighttime deposition events. The chamber flushing rate appears to have a very significant impact on the measured fluxes and on the response behavior to environmental change. This study demonstrates that both experimental differences and small-scale regional variability introduce large uncertainty in the estimation of natural Hg air-surface exchange by different flux chamber techniques. Also, the impact of environmental parameters on Hg air-surface exchange was studied. Rain events led to a strong increase in the Hg emissions, even when the covered soil remained dry, suggesting that the apparent chamber footprint is larger than the actually covered area. Exclusion of sunlight led to decreases in Hg emissions. Statistical analysis revealed the strongest correlations between the measured Hg fluxes and radiation and wind speed. Weaker correlations were observed with air and soil temperature and wind direction (probably due to local Hg sources). Fluxes were also inversely correlated with relative humidity. ¿ 1999 American Geophysical Union