Interdisciplinary Programs Office Division of Environment and Sustainability 240 Air Quality Data Analysis and Interpretation Supervisor: GU Dasa / ENVR Student: WAH Monique Zoe / SSCI Course: UROP1000, Summer GEO-KOMPSAT-2A (GK2A) is a geostationary satellite located at 128.2°E. Chlorophyll-a concentration can be calculated from GK2A data following Murakami’s method illustrated in the paper “Ocean Color Estimation by Himawari-8/AHI”. GK2A visible band data will first be displayed and merged as a true color image. GK2A’s Advanced Meteorological Imager (AMI) spectral response function will be plotted and analyzed for simulation of remote-sensing reflectance value. During the calculation process, Inherent Optical Property Model and Band-ratio Algorithm are utilized. NOMAD data is also used for deriving the coefficients of Band-ratio Algorithm. Chlorophyll-a concentration is displayed as a color map, and comparison between calculation result with MODIS and Himawari 8 product is present in the paper. Air Quality Data Analysis and Interpretation Supervisor: GU Dasa / ENVR Student: WANG Yiyi / CHEM Course: UROP4100, Summer I reported on the sensitivity of ground surface ozone to the changes on the fluctuations in concentrations of volatile organic compound (VOC) and nitrogen oxides (NOx = NO + NO2) observed in a windless dry sunny day at the coast of Silver Strand Bay, using a 0-D box model which has the Master Chemical Mechanism implemented for simulation as well as Python for time series plotting and correlation analysis. The findings of this study educated me about VOCs and their impact on O3 formation, and will be helpful in formulating emission control strategies for coping with O3 pollution related to photochemical air pollution in Hong Kong. Use WiFi Technologies to Predict Building Energy Consumption Supervisor: LU Zhongming / ENVR Co-supervisor: QU Huamin / CSE Student: CHAWLA Anhad Singh / SUEE Course: UROP1100, Fall This research report explores the relation between several variables that predict energy usage by creating machine learning regression models. Data for new variables regarding building use type was collected using Google Maps and Street View and added to the existing database. The code was adjusted to accommodate the new data and after some data cleaning, 13 models used the data to create a predictive model. The specially curated database is discussed in detail with an explanation for the variables chosen for model input. The efficacy of the models is discussed and future research directions are suggested.
RkJQdWJsaXNoZXIy NDk5Njg=