Science Research

My long-term goal is to help improve people’s lives by advancing the field of meteorology. We see evidence daily that the world is facing greater threats from climate change. Advanced weather models based on machine learning, like the one I’m currently building, will become integral to our success as a global community.
​
​The 2020 hurricane season was catastrophic for Central America and the Gulf Coast. Because the conditions that caused this hyperactivity will only become more frequent, my research holds the potential to help communities and save lives.

Research portfolio

​2021
Meteorology - Year Four

Predictive Modeling of Tropical Cyclone Rapid Intensification by Analyzing Convective Patterns with Convolutional Neural Networks

• ​1st Place, Senior Earth & Environmental Science, 2021 Brevard Intracoastal Science Fair
• 1st Place, Senior Earth & Environmental Science, 2021 Lockheed Martin Science Challenge
• 1st Place, Senior Earth & Environmental Science, 2021 State Science & Engineering Fair of Florida (SSEF)
• Best-in-Fair Grand Award - Ying Scholar for Physical Sciences, 2021 SSEF
• Office of Naval Research Science Award
• International Science & Engineering Fair (ISEF) finalist in May, 2021

ABSTRACT
​Rapid Intensification (RI), a 30+ kt increase over a 24-hour period in a storm’s 1-minute maximum sustained winds, is a potentially catastrophic event occurring within tropical cyclones. The leading current RI model has an accuracy rate of 82% and analyzes large-scale environmental trends, such as surrounding sea surface temperatures, but ignores inner-core convective patterns within storms. Accuracy could therefore potentially be improved by analyzing convective patterns. 

A Convolutional Neural Network (CNN) is a machine learning program specializing in image analysis. CNNs work by taking local convolutions from images, running it through several layers of neurons, where convolutions are compared, until an output pattern can be determined. By using a CNN to analyze GOES GridSat satellite images, filtered using infrared radiation to detect cloud top temperatures, convective patterns could emerge. To achieve this, a CNN was built in Python using package Tensorflow, and run on storm images classified as RI or not RI. This model was tuned to focus on four different “crops” of data at different radii around a storm’s center. 

The storm’s eyewall crop saw a 10th-epoch mean validation accuracy of 77.8%. This indicates that the model was 77.8% accurate at detecting rapid intensification in storms - only 4.2% lower than the SHIPS RI model. In addition, validation accuracy increased as the crop became more focused on a storm’s eyewall. By integrating more datasets, potentially those containing similar variables to those studied in SHIPS RI, accuracy could be increased.
​

​2020
Meteorology - Year Three

Continuation study on the effects of the Saharan Air Layer (SAL) on Atlantic tropical cyclones

• 1st Place, Senior Earth & Environmental Science, Brevard Intracoastal Science Fair
• Invitation to compete, 2020 Florida State Science & Engineering Fair (canceled due to COVID-19)
• American Meteorological Society Award
• NASA Earth Systems Award

ABSTRACT
​This project assesses the relationship between the Saharan Air Layer (SAL) and tropical cyclones in the North Atlantic basin. The SAL is a mass of dust that blows off the Sahara Desert every three to five days over the tropical Atlantic. It brings with it strong (30+ kt) vertical wind shear and extremely dry air, both of which have negative impacts on tropical cyclones. This research is relevant because an improved understanding of this interaction can improve cyclone forecasts, potentially saving lives.

To compare this relationship, the average strength of the SAL as measured by the RGB values in SAL-filtered satellite imagery was taken in four quadrants surrounding a storm. These RGB values were then calculated into an SAL Index. SAL Index values were then compared to the intensity of tropical cyclones from 2005 to 2011 (as measured by maximum windspeed, minimum central pressure, and Accumulated Cyclone Energy (ACE)) in a direction-based and overall spatial average. The CIMSS SAL Intensity Database and the NOAA HURDAT Tropical Cyclone Database were the primary sources of data used to analyze this relationship.

The research demonstrated that strong SAL conditions did not correlate to reduced tropical cyclone intensity, measured by windspeed and pressure. This may be due to the negative effects of the SAL being offset by increased atmospheric temperature differentials, increased mid-level moisture, and excellent upper-level outflow pathways surrounding an SAL burst. Overall, the strongest SAL conditions surrounding and in front of storms generally corresponded to the most consistently strong tropical cyclones as measured by ACE.

​2019
​Meteorology - Year Two

The effects of the Saharan Air Layer (SAL) on Atlantic tropical cyclones

• 2nd Place, Senior Earth & Environmental Science; 2019 Brevard Intracoastal Science Fair
• NOAA Award
• Office of Naval Research Award
• American Meteorological Society Award
• NASA Earth Systems Award

ABSTRACT
​This project assesses the effects of the Saharan Air Layer (SAL) on the intensity of tropical cyclones in the North Atlantic Basin. Statistics indicate that the SAL has a negative impact on hurricanes, as it can disrupt deep convection necessary to maintain storm clouds, and brings with it strong wind shear that blows away cloud tops. This research is relevant because it is progress towards more accurate prediction of hurricane intensity, which can save lives and minimize property damage.

To compare this relationship, the strength of the SAL as measured by the RGB values in SAL-filtered satellite imagery was taken at a point one day ahead of a tropical cyclone's path. These RGB values were then calculated into a SAL Index. SAL Index values were then compared to the intensity of tropical cyclones from 2004-2006. The CIMSS SAL Intensity Database and the NOAA HURDAT Tropical Cyclone Database were the primary sources of data used to analyze this relationship.

The research demonstrated that periods and locations sampled with strong SAL conditions did not strongly correlate to reduced tropical cyclone intensity, measured by  maximum wind speed and minimum central pressure. This may have been due to local variances in SAL events, corresponding to different SAL Index values across larger areas and altitudes. These local variances can foster increased sea surface warmth, excellent outflow pathways, and mid-level moisture, all of which may contribute to tropical cyclone intensification. Overall, the strongest SAL conditions generally corresponded to the most consistently strong tropical cyclones.

2018
Meteorology - Year One

The effects of the El Niño Southern Oscillation (ENSO) on Atlantic tropical cyclones

• 2nd Place, Senior Earth & Environmental Science; 2018 Brevard Intracoastal Science Fair
• American Meteorological Society Award
• NASA Earth Systems Award

ABSTRACT
This project assesses the effects of different phases of the El Niño Southern Oscillation (ENSO) on the frequency and intensity of Atlantic tropical cyclones. Statistics  indicate that ENSO has an effect on these variables. The research is relevant because more accurate prediction of storms will help people  be better prepared, potentially minimizing loss of life and property damage.

To compare these variables, the strength of the ENSO as described by the Oceanic Nino Index (ONI) three-month mean values was looked at at the time of each named North Atlantic tropical hurricane from 2008-2017. The ONI three-month means were then compared to the strength of named storms as measured by minimum pressure and maximum wind speed. In addition, the yearly overall strength of the ENSO was compared to the quantity of cyclones that developed in the Atlantic Ocean. The NHC Data Archive and Microsoft Excel were the primary data source and tool used to analyze the relationship.

The research showed that either a moderately strong El Niño or a moderately strong La Niña were concurrent with the most intense storms from 2008-2017. This most likely happened because moderately strong La Niñas are associated with lower wind shear (a limiting factor on Atlantic tropical cyclones) and moderately strong El Niños bring higher water temperatures in the Main Development Region (MDR) of the Atlantic Ocean. These factors favored the development of more and stronger storms during moderate El Niños.

​2017
Microbiology - Year Two

The effects of unfiltered and UV-filtered solar radiation on gram-positive and
​gram-negative bacteria

• 1st Place, Junior Microbiology; 2017 Brevard Intracoastal Science Fair
• 2nd Place, Junior Microbiology; 2017 Florida State Science & Engineering Fair

​ABSTRACT
​This project studies the effects of unfiltered and UV-filtered solar radiation on gram-negative E. coli K-12 and gram-positive M. luteus bacteria. I am researching effective and sustainable ways to solve the problem of bacterial infection.

​In my Year 1 project, I found that solar radiation is effective at sterilizing E. coli bacteria. This year, I studied what spectrum of sunlight is most effective, and whether the sunlight sterilization method would work on both gram-positive and gram-negative bacteria. I tested this by placing inoculated agar plates in direct sunlight, some with a UV-filter film cover, and some without. Every 30 minutes, I removed agar plates from each group. The plates were then incubated so that colony forming units (CFUs) would develop, which I then counted.

I found that the UV-filter made a small difference in the survival rates of E. coli K-12, which became sterilized after 1 hour under direct sunlight and 1½ hours under the UV-filter. M. luteus survived longer under the filter, becoming sterile after 30 minutes in direct sunlight and 2 hours under the filter.

I think this happened because M. luteus is a non-enteric bacterium. E. coli K-12 is enteric, meaning it lives inside animal intestines, and therefore is less exposed to sunlight naturally. I believe the filter helped M. luteus’ survival by reducing the UV intensity in the solar radiation present. I conclude that solar radiation is an effective and sustainable tool for sterilizing bacteria, and for helping make food and water safe for human consumption.

2016
​Microbiology - Year One

The effects of radiation on the growth of
​E. coli K-12 bacteria

• 1st Place, Junior Microbiology; 2016 Brevard Intracoastal Science Fair
• 5th Place, Junior Microbiology; 2017 Florida State Science & Engineering Fair
• Doctor's Goodwill Foundation Award

ABSTRACT
This project studies the effects of different forms of radiation on the E. coli K-12 strain of bacteria. I chose this experiment to try to find an inexpensive, efficient and sustainable method of sterilizing the harmful E. coli bacteria.

To accomplish this, I diluted the bacteria to a 1 part bacteria to every 800,000 parts water dilution. I then put drops of this solution onto agar plates, and irradiated the plates using cheap and accessible forms of radiation, such as the sun. I then removed the plates from exposure to the radiation at regular two-hour intervals, and incubated them. The next day, I counted the numbers of CFUs that developed on the agar plates.

I found that the sun was the most effective at killing the bacteria, while the other radiation sources showed relatively stable and consistent trends. After four hours, the bacteria in the solar-irradiated plates had disappeared, going from an average of 62 CFUs per plate to zero by the six-hour sample.

I think this happened because of the UV-B radiations in the sunlight, which effectively poison the bacteria. Because of this, I conclude that the sun and its potentially harmful radiation is an effective tool in the process of sterilizing E. coli, making food safe for consumption.​