A novel portable device for the direct measurement of alveolar breath carbon monoxide (CO) was developed. The major components of the device include a mouthpiece, non-rebreathing two-way valve, Teflon tube, and CO dosimeter. An alveolar CO measurement can be completed within 1.5 min when using the proposed device and measurement protocol. Measurements could be read to the nearest 0.1 ppm. Humidity did not influence the CO measurements taken by the CO dosimeter, plus there were no problems associated with the recovery and carryover of CO through the device. The criterion for significance in statistical analyses was p< 0.05. The average recovery was 103 and 99% for recovery and carryover experiments, respectively. Test results using the proposed alveolar CO measurement system reflected a good reproducibility. This reproducibility was also supported by the finding that the relative standard deviations (RSDs) of the data sets were less than 7％ for the loss experiment and less than 8% for the carryover experiment. Consequently, it would appear that the proposed device can be effectively applied to measure CO levels found in breath, thereby overcoming several disadvantages associated with the conventional bag and adsorbent tube sampling methods.

The pH, electric conductivity. and the major ionic components were analyzed for the precipitation samples collected at 1100 Site of Mt. Halla and Jeju city. The quality of analytical data was verified by the comparison of ion balances, conductivities and acid fractions, all of which correlation coefficients were over 0.952. The ionic strengths lower than 10$^{-4}$ M were found in 57 and 28% at 1100 Site and Jeju city respectively. The precipitation in Jeju city was influenced more by the oceanic effect than those in 1100 Site. The acidification of precipitation was caused mostly by S $O_4$$^{2-}$and N $O_3$$^{[-10]}$ in both areas, and the organic acids have contributed to the acidity with only 7~8%. The neutralization factors by N $H_3$ were about 44 and 47％ at the 1100 site and the Jeju city, respectively, whereas those by CaC $O_3$were 21 and 24%, and the free acidity were about 38 and 28％ at two sites. From the investigation of seawater and soil enrichment factors, the S $O_4$$^{2-}$, N $O_3$$^{[-10]}$ and N $E_4$$^{+}$ were immigrated by other sources rather than from the seawater or soil origins. but not in the case of $Mg^{2+}$, C $l^{[-10]}$ , N $a^{+}$, and $K^{+}$. Factor analysis has shown that the precipitation at the 1100 site had been influenced mostly by anthropogenic sources, followed by soil and sea-water sources. On the other hand, the precipitation at the Jeju city was mainly influenced by oceanic sources, followed by anthropogenic and soil sources.urces.

Receptor modeling is one of statistical methods to achieve reasonable air pollution strategies. In order to maintain and manage ambient air quality, it is necessary to identify sources and to apportion its sources for ambient particulate matters. The main purpose of the study was to survey seasonal trends of inorganic elements in the coarse and fine particles. Second, this study has attempted emission sources qualitatively by a receptor method, the PMF mo-del. After that. both PMF (positive matrix factorization) model and TTFA (target transformation factor analysis) model were applied to compare and to estimate mass contribution of coarse and fine particle sources at the receptor. A total of 138 sets of samples was collected from 1989 to 1996 by a low volume cascade impactor with 9 size fraction stages at Kyung Hee University in Korea. Sixteen chemical species (Si, Ca, Fe, K, Pb, Na, Zn, Mg, Ba, Ni, V, Mn, Cr, Br, Cu. Co) were characterized by XRF. The study result showed that the weighted arithmetic mean of coarse and fine particles were 51.3 and 54.4 $\mu\textrm{g}$/㎥, respectively. Contribution of both particle fractions were esti-mated using TTFA and PMF models. The number of estimated sources was seven according to TTFA model and 8 according to PMF model. Comparison of TTFA and PMF revealed that both methodologies exhibited similar trends in their contribution pattern. However, large differences between contributions were observed in some sour-ces. The results of this study may help to suggest control strategies in local countries where known source profiles do not exist.

The real-time modeling system, named AirWatch System, has been developed to evaluate the environmental impact from a large source. It consists of stack TMS (TeleMetering System) that measures the emission data from the source, AWS (Automatic Weather Station) that monitors the weather data and computer system with the dispersion modeling software. The modeling theories used in the system are Gaussian plume and puff models. The Gaussian plume model is used for the dispersion in the simple terrain with a point meteorological data while the puff model is for the dispersion in complex terrain with three dimensional wind fields. The AirWatch System predicts the impact of the emitted pollutants from the large source on the near-by environment on the real -time base and the alarm is issued to control the emission rate if the calculated concentrations exceed the modeling significance level.