This study estimates the correlation dimensions of the land-sea breeze phenomenon, that has a clear diurnal cycle, in order to gain a more detailed understanding of this phenomenon. The data adopted include north-south wind velocity component(v) and temperature(T) time series that were observed at Kimhae Airport and Inje University over a period of 5 days, from the 4th to the 8th of August, 1994. The embedding phase space of the time series were reconstructed from 2 to 14 dimensions, and the correlation dimensions of the attractors were then estimated. The results show that the land-sea breeze phenomenon exhibits a deterministic chaos with non-integer correlation dimension values between 2 and 3. Accordingly, 3 is the minimum number of independent variables required to model the dynamics of the landsea breeze phenomenon in the Kimhae area. Since the saturated embedding dimension, when the correlation dimension remains unchanged, is larger for the wind velocity v-component than for temperature, this indicates that wind velocity is susceptible to topology.

According to numerical experiments, the Sokcho Eddy is produced at $37 5~39.0^{\circ}N$ by strong offshore winds, whereas the Ulleung Eddy is produced at $35~37^{\circ}N$ by an inflow variation of the Tsushima Current. These locations compare well with visual observations. The nonlinear 1$1/2$-layer model showed that most of the East Korea Warm Current (EKWC) driven by the Tsushima Current form the Ulleung Eddy that is larger and stronger than the Sokcho Eddy. In contrast, the nonlinear 2$1/2$-layer model showed that most of the EKWC travels further northward due to a strong subsurface current, thereby enhancing the Sokcho Eddy making it larger and stronger than the Ulleung Eddy. The Sokcho Eddy is also produced relatively offshore due to an eastward subsurface current in the frontal region. Using the 1$1/2$-layer model, when the mass of the Tsushima Current decreases, the two eddies are weakened and produce a circular shape. In the 2$1/2$-layer model the EKWC pushes the Ulleung Eddy northward after 330 days, next the Sokcho and Ulleung eddies begin to interact with each other, and then after 360 days the Ulleung Eddy finally disappears absorbed by the relatively stronger Sokcho Eddy. This behavior compares favorably with other visual observations.

Density and temeprature fronts are common features of the ocean. However, frontal dynamics are not quasi-geostrophic because the isopycnal deflections associated with fronts are large compared with the scale height of the hydrostatic geopotential. The frontal geostrophic model, developed by Cushman-Roisin et al.(1992) is generally used fro describing the dynamics of surface-density ocean fronts, whereas the two-layer frontal geostrophic model is used for fronts on a sloping continental shelf. This paper investigates the baroclinic nonlinear stability of surface-density ocean fronts and fronts on a sloping continental shelf using the two-layer frontal geostrophic model mentioned above. Nonlinear stability criteria for the two kinds of fronts are obtained using Arnol'd's (1965; 1969) variational principle and a prior estimate method. This is the first time a nonlinear stability criterion for surface ocean fronts has been established, furthermore, the results obtained for fronts on a sloping bottom are superior to any previous ones.

The quality of the effluent from an activated sludge aeration tank can deteriorate when the substrate removal rate decreases due to an abrupt reduction in the DO concentration, which is affected by such operating conditions as the loading rate, temperature, wastewater composition, and so on. In this research, a DO control system that includes a PI (proportional-integral) controller/Hiraoka controller was developed and applied to a pilot-scale activated sludge process, then its acceptability was estimated. The applicability of the respiration rate to DO control was also estimated. The respiration rate indicated a variety of input organic loading rates, which is the main disturbance to the DO concentration in an aeration tank. When the influent concentration incrementally decreased and increased between CODcr 1,000 mg/l and 100 mg/l, the control system with a PI controller exhibited a good llperformance-the average DO concentrations were 2.00$\pm$0.14 mg/l and 1.88$\pm$0.15 mg/l (set value was 2.0 mg/l), respectively, and the settling time was just 10 minites. When the control system was operated for 4 days, the DO concentration was 1.99$\pm$0.18 mg/l and 32.6% of the air flowrate was saved. However, the fluctuations in the respiration rates and air flowrates were severe, which could be harmful to the stability of the biomass and mechanical stability of the blower. A possible approach to solve this problem may be the simultaneous control of the loading rate and DO concentration.

Photosynthesis and transpiration rates were simultaneously measured in attached sunflower leaves(Helianthus annuusL. cv. Russian Mammoth) during exposure to $NO_2$ and $O_3$ to determine the effect of mixed gan on photosynthesis and the stomatal aperture. The application of $O_3$ alone reduced both the net photosynthetic and transpiration rates. An analysis of the $CO_2$ diffusive resistances indicated that the main cause affecting photosynthesis reduction during $O_3$ exposure was not the internal gas phase of the leaf $(rCO_2^{liq})$ but rather the liquid phase or mesophyll diffusive resistance $(rCO_2^{liq})$, suggesting that there is a very concomitant relation between photosynthetic reduction and $rCO_2^{liq}$. The application of NO2 alone caused a marked reduction of the net photosynthesis yet no significant reduction of transpiration, indicating that NO2 affects the $CO_2$ fixation processes with no inluence on the stomatal aperture. A greter reduction in the photosynthesis of sunflower plants was caused by the application of $NO_2$ alone as compared to a combination of $NO_2$ and $O_3$. $NO_2$ alone reduced the photosynthetic rate by 90%, whereas a mixture of NO2 and O3 reduced it by 50%.

The solubility of H2S is an important parameter for modeling H2S absorption. Since the direct measurement of H2S solubility in aqueous amine solutions is impossible, this work initially attempted to develop an H2S analogy method, however, this was unsuccessful. Consequently, H2S solubilities were measured in aqueous amines which were completely protonated with HCI over a temperature range of 25-60℃. The solvents investigated in this work included 0-50% aqueous solutions of methyldiethanolamine and diethanolamine. Thereafter, a new empirical correlation was developed that can predict Henry's constant for H2S using only the solubility of H2S in water and t도 molecular weight of the aqueous solvent.

One of the most important parameters required to model the absorption of CO2 into aqueous alkanolamine solutions is physical solubility. However, since CO2 reacts in amine solutions, its physical solubility cannot be measured directly. As a result, a nonreacting gas which is similar to CO2 has to be used such as N2O. The solubility of nitrous oxide (N2O) in aqueous solutions of 0wt%-50wt% MDEA, 0wt%-30wt% DEA, and 50wt % total amine with DEA/MDEA molar ratios of 0.05, 0.25, 0.5 and 0.67 was measured using a modified Zipperclave reactor over a temperature range of 293-353 K with near atmospheric partial pressures of N2O. the solubility data from this work were K with near atmospheric partial pressures of N2O. The solubility data from this work were found to be in good agreement with previously reported data where available.