NASA/NOAA FLOOD MAPPING PROJECT
This study investigates an approach to map 3D flood map (i.e., floodwater extent and depth) using UAV high-resolution imagery and LiDAR for Hurricane Matthew. We utilize a deep learning approach to map flooded area extent from post-event UAV images, and then employ spatial statistics to estimate the water level depth of the flooded areas leveraging on the corresponding DEM. Afterward, a built-up surface auxiliary dataset was combined with the generated flood depth result to map and analyze flood impact risk in settlement areas within the Town of Grifton, NC during Hurricane Matthew. Our results showed that settlement areas in Grifton exhibit different risk levels from a 3D flood depth perspective. This information could significantly enhance near real-time emergency response strategies, as well as future adaptation and mitigation initiatives.
The study results showed a maximum flood water depth of 8ft with an RMSE score of 0.88. The model utilized for estimating the floodwater extent generated an overall accuracy of 97%.
Estimated floodwater extent.
Estimated 3D flood risk map depth with building footprint within Grifton settlement area.
Estimated 3D flood water level depth with building footprint in Grifton.
MAPPING HISTORY
PROJECT SUMMARY
This was a collaborative project between my lab and the Political Science and History Department - North Carolina A&T State University, led by Dr. Arwin Smallwood, the department Chair. I conducted this proximity analysis as part of the project to investigate and answer the question; Which plantation farms are most vulnerable to the infestation threat from surrounding swamp areas?
To answer the question, the proximity analysis was conducted to classify the farms into three main Risk assessment categories, based on their proximity to the swamp areas.
Risk assessment Classes;
High Risk: Farm plantations, closest to the swamps.
Neutral: Farm plantations that follow the closest farms to the swamps.
Low Risk: Farm plantations, further away from the swamps
Farm planation infestation risk map
Statistics chart showing the number of farm plantations at each risk level. 39% of the total farm plantations (159) within the indianwoods territory were highly at risk of the swamp infestations.
MAPPING COMMUNITY GREEN INITIATIVES
As an intern, I worked with 7 community green groups in different neighborhoods within the city of New Haven. I Collaborated with these community groups as an intern for URI throughout the summer to implement the green initiatives of the respective groups. These initiatives mainly involved the planting of trees, shrubs, and perennials. I worked with other groups to also create walk paths and beds in public parks such as Fairmont Park and Kensington Park within the Fair Haven, and Dixwell neighborhoods, respectively. In all, I planted about 20 plants with all the groups.
I then utilized the ArcGIS Pro, and field map to collect data and map all the implemented green initiatives for each of the groups.
PROJECT SUMMARY
Friends of Kensington - Kensington Park, Dixwell-New Haven
Friends of Fairmont Park - Fairmont Park, Fair Haven Heights - New Haven
Friends of Mill River Trail - Mill River Trail, Fair Haven -New Haven
Watson & Basset Community Greenspace- Watson & Basset st., New Haven
Oyster Point Community Greenspace - City Point, New Haven
DEVELOPMENT TREND ANALYSIS
The City of New Haven is an old mid-sized industrial city experiencing a revamp of development. A revitalized development trajectory is noted to have started from Mayor Richard Lee’s era in the 1950s and 1960s. This development is primarily regulated by agencies whose duties and activities are enshrined in the Zoning Ordinance of the City; originally published in 1995. The Ordinance has since been regularly updated on an annual basis to help regulate the sustainable growth and development of the city. The City of New Haven is currently fostering green developments, and this initiative requires an updated and well-integrated Green Ordinance to better regulate and direct development projects and the appropriate agencies in charge. To do this, reviewing the development trajectory and regulatory actions by agencies based on existing data could significantly inform such green development initiatives within the city. Therefore, this study utilized the existing application databases of the City Plan Commission (CPC), and the Board of Zoning Appeals (BZA) to identify and analyze development trends at the neighborhood level within the city over a period.
The study aimed to answer the following questions;
• Where are new developments happening within the city?
• Are there areas of concentration for specific types of development?
• How do such patterns impact the city’s green ordinance?
PROJECT SUMMARY
Time series map showing counts of special exceptional approvals per neighbourhood
Time series map showing counts of site plan approvals per neighbourhood
Map showing counts of development project applications per census block
NASA ENVIRONMENTAL JUSTICE
This project involved characterizing and monitoring mult-decadal urban heat and tree canopy cover for the state of Connecticut and the city of New Haven. The project utilizes satellite and aerial images (MODIS, LANDSAT, NAIP) as well as socio-economic data to develop geospatial information to inform decision-making at the state and city level. As part of a team of 3 research assistants, I ;
Conduct a block-level analysis of street trees within neighborhoods in the city of New Haven.
Estimated the mean Temperature and NDVI from Landsat imagery and did a block-level comparative analysis with buildings/impervious surfaces per census block in the City of New Haven.
PROJECT SUMMARY
A
B
Map A shows a block-level comparative analysis of the number of street trees and building counts per census block. This was to identify whether census blocks with high building counts have more street trees. Map B also shows building and mean temperature (avg. 2022) per census block. This was to identify if blocks with higher building counts have a higher mean surface temperature compared to the surrounding blocks.
This map show building counts per block in the city of New Haven. This is to help identify blocks with higher impervious surfaces or concentration of activities.
URBAN AFRICA PROJECT
The Urban Africa project involved a time series mapping of the horizontal and vertical dimensions of building infrastructures within urban areas in Africa, using the Landsat image collection and Convolutional Neural Network algorithms (python Tensorflow). The study focused on all urban areas across Africa. I was part of a team of 5, led by one postgraduate and one P.I (Principal Investigator ).
As part of the team, I performed the following tasks and analysis;
Conducted a literature Review of existing methods for mapping horizontal and vertical dimensions of building infrastructures.
Reviewed existing training datasets such as the Local Climate Zone(LCZ) dataset, using python and ArcGIS Pro to assess how they could be useful to the project.
Extracted and processed 60,000 open street map buildings across 6 African Cities using QGIS, and Google Earth plug-in.
Utilized Arcpy and R to process and convert building footprints data into label mask(raster) to train the CNN algorithms.
Downloaded Landsat Image entire Collection using Google Earth engine scripts.
This map shows the total number of available LCZ training datasets per country. This helped to focus our effort on countries with low or no available datasets.
This is a sample training site for Accra. It constitute about 15,000 building footprints(left image).I processed the data into a raster label mask(right image) using Arcpy and R scripts.
openstreet map building footprint in 30 sq.m grids
Raster mask showing building area coverage per 30 sq.m
REAL-TIME SPATIAL DATA COLLECTION & MONITORING
I utilized ArcGIS Pro, ArcGIS Online, and ArcGIS field map app to design and collect field data for my research paper. I used arcGIS Pro preprocess and process shapefiles such as building footprints and polygon boundaries for my area of interest.
I then uploaded them to the ArcGIS cloud repository and developed a web map. Afterward, I used the web map to develop a field app using the ArcGIS field map. The purpose of the fieldwork was to collect attribute data of about 600 building structures within the three metropolitan areas in Ghana. The building attributes collected included; building height(based on floor numbers)roofing materials, wall materials, etc. These are ground truth data for my research paper. Designing and developing the field map improved the efficiency of my fieldwork, and immensely aided the organization of the collected data.
FIELD DATA DESIGN AND DATA COLLECTION
ArcGIS Field map app
I also developed an information dashboard to monitor the field data collection in real time. This helped to monitor my activities on the field and also present preliminary statistics and analysis of the field data collected in real time. I utilized ArcGIS online and the ArcGIS information dashboard app.
REAL-TIME MONITORING OF FIELD WORK
Interactive Real-time monitoring information dashboard
scan
OTHER PROJECTS
IMAGE PROCESSING AND ANALYSIS
I estimated the Building Density composition for Metropolitan Areas in Ghana. The Building Density was calculated as the coverage of buildings per 30 square meters Pixel. I extracted the building density from Landsat Image Collection using deep learning Learning and Remote Sensing/GIS techniques. The purpose of this analysis was to produce statistics and analysis of the physical characteristics of the built environment within Ghana’s Metropolitan areas. This was to inform and foster sustainable and resilient decision-making towards sustainable urban development.
Building Density Analysis
3D visualization of building density distribution for 5 cities (Using ArcScene).
SPATIAL ANALYSIS
I utilized the lidar 3D point cloud datasets to estimate building heights. This was part of an Advanced Geospatial Analysis class project. I processed the point cloud datasets using ArcGIS pro to estimate building heights at North Carolina A&T State University's main campus.
Lidar point cloud analysis
Lidar 3D points cloud for all features
Lidar 3D points cloud for buildings
I did this spatial analysis as a GIS Officer at the Ghana Statistical Service. It was to help identify Enumeration areas, whose building structures will be hard to enumerate as part of the 2021 Population and Housing Census. Here, three spatial datasets; building footprints, roads, and population count for each enumeration area were analyzed and combined to generate a final map using ArcGIS Pro. The combined map shows the levels of difficulty for each enumeration area for the Housing Census.
So enumeration areas with high building and population counts, and low clusters of roads, were hard to enumerate.
This Analysis helped to allocate more resources to enumeration areas projected to face difficulties in the 2021 the Population and Housing Census activities.
Hard to Count Analysis
Distance and Emergency Response Time analysis for Fire Trucks
I conducted these analyses as an RS/GIS lab project in collaboration with the Ghana Fire Service.
The distance analysis map shows the various accessible routes that could be taken by Firefighters during emergencies. The analysis is based on the Katamanto Market Center, Accra which is highly prevalent for fire incidents. The distance analysis helped to identify the shortest access routes that could be taken from the nearest fire stations to the market, in case of fire emergencies.
Distance Analysis
Distance map analysis
The response time analysis displays the average response time of all fire stations within the Accra Metropolitan Area(2019). The average response time data of each fire station were collected from the fire stations; they were then used to calculate the average distance of each fire station; then a buffer of the average distance for each fire station were created to show the distance coverage of each fire station. It was then compared with the National Fire Protection Association(NFPA)’s standard distance coverage of a fire station. The analysis, as displayed in the map, showed that, all the fire stations, at the time of the analysis, covered more distance than the standardized distance.
Response Time Analysis
Response Time map