Project Repository

https://github.com/samtduke/TolbertAndinoPantojaBarbera_ENV872_EDA_FinalProject

0.0.1 List of Tables

Dataset Information

Seven-Day Time Series Analysis: Trend Patterns

0.0.2 List of Figures

Time Series Analysis: Daily and Historical Trend Patterns

Seven-Day Time Series Analysis: Trend Patterns

0.1 Context and Motivation

Hurricane Helene struck North Carolina in late September, rapidly intensifying to a Category 4 storm with sustained winds of 140 mph (Thiem & Lindsey, 2024). The hurricane brought catastrophic rainfall, widespread flooding, landslides, destructive winds, and tornados, causing severe disruptions to both the economy and daily life (Cooper, 2024;Thiem & Lindsey, 2024). Months later, the state is still recovering from the storm’s far-reaching effects. Particularly striking is that Helene followed closely on the heels of Hurricane Milton, a Category 5 storm that had already devastated the west coast of the United States. These concurrent events underscore the increasing severity and frequency of hurricanes, highlighting the urgent need to better understand the mechanisms behind and the impacts of such extreme weather events.

The devastation caused by Hurricane Helene has left a profound and lasting impact on North Carolina(Cooper, 2024). The storm wreaked unprecedented destruction, leaving thousands of homes either destroyed or severely damaged. Millions of residents were cut off from critical infrastructure, including water, electricity, telecommunications, and healthcare services (OSBM, 2024). Key infrastructure, such as roadways and bridges, was also heavily damaged. Economically, the storm’s effects will be felt for years to come. Preliminary estimates of the damage and recovery needs across the state total approximately $53.6 billion, which includes $41.1 billion in direct damages, $7.6 billion in indirect impacts, and $4.8 billion needed for strengthening and mitigation efforts (Cooper, 2024; OSBM, 2024). These figures highlight the enormity of the storm’s impact compared to other recent hurricanes, such as Hurricane Florence in 2018, which resulted in $16.7 billion in damages(NC Gov., 2018). Beyond the immediate economic losses, the storm significantly altered the state’s physical landscape, eroding valuable topsoil, changing river courses, and submerging homes that had stood for generations. The long-term effects on agriculture and other industries will continue to be felt for years to come (OSBM, 2024).

Why should we understand rainfall trends and why is disaster preparedness important?

The rainfall from Hurricane Helene was both unprecedented and catastrophic. In some areas of North Carolina, over 26 inches of rain fell within a 24-hour period, resulting in widespread flooding that devastated entire towns across 27 counties (OSBM, 2024). Analyzing precipitation patterns and discharge before, during, and after the hurricane is crucial for understanding its long-term effects on regional climates and water systems. This information is vital for informed decision-making in disaster preparedness and water resource management.

Hurricane Helene serves as a reminder as a need to strengthen communities against flooding and extreme weather events. Climate change knows no boundaries, and we should not be surprised when such extreme events continue to occur. By focusing on Helene and its unprecedented nature, this study establishes important benchmarks, analyses, and comparisons that will be invaluable for future disaster response planning.

Analyzing precipitation patterns can also enhance the resilience of city and town infrastructure. Accurate forecasting of precipitation trends is vital for designing and maintaining stormwater management systems that safeguard cities, communities, and municipalities from the impacts of climate-induced rainfall.

This study examines rainfall trends associated with Hurricane Helene, emphasizing the historical significance as a rare event with far-reaching consequences. Understanding these patterns offers a foundation for contextualizing the storm’s impact within the broader context of climate change, while also identifying anomalies in long-term climatic trends.

0.2 Research Questions

Question 1: What were the temporal and spatial rainfall patterns in Asheville, Sugar Grove, and Pigeon River following Hurricane Helene?

Question 2: How do these three areas experience rainfall and flooding impacts differently?

Question 3: How have these three areas historically experienced rainfall and flooding impacts?

0.3 Data Description

To assess the impact and severity of Hurricane Helene, we analyzed data from the United States Geological Survey (USGS). This dataset highly valuable for understanding extreme weather events, given its accuracy and consistent daily collection. USGS stream gages, which record daily discharge volumes in cubic feet per second, were selected for their reliability and ability to capture critical hydrological changes during significant events. The USGS dataset also has immense historical data records that are valuable in understanding trends over time.

Our analysis centered on daily precipitation trends, flood data, and hurricane patterns. To capture the diverse geographical and hydrological impacts of Helene across Western North Carolina, we selected three focal sites: Sugar Grove, Asheville, and Pigeon River. These locations were strategically chosen based on their proximity to Helene’s storm path, providing a comprehensive representation of the storm’s effects across Western North Carolina. We specifically looked at the French Broad River but will be labeling it as Asheville throughout our study.

The study examined three critical parameters: precipitation, stream discharge, and gage height. For each parameter, we calculated daily means and cumulative sums to identify patterns and assess the storm’s severity across different regions.

Sugar Grove: Selected for its agricultural setting and the significant impact it experienced from Hurricane Helene.

Asheville: Chosen due to its size and role as a metropolitan hub in rural Western North Carolina.

Pigeon River: Included for its proximity to the Tennessee border, highlighting the need for cross-border storm management strategies.

Together, these datasets and locations provide a multi-dimensional view of Hurricane Helene’s impact. This approach bridges historical context, storm-specific data, and hydrological consequences, ensuring that the study not only quantifies the immediate effects of the hurricane but also contributes to long-term disaster preparedness and climate resilience.

USGS stream gage data includes the latitude and longitude of the gage, discharge in cubic feet per second, and the date and time of data recording.

Table 1: Dataset Structure for Analyzing Hurricane Helene’s Impact
Variable Name Description Data Type
Location Name of the site (e.g., Asheville, Sugar Grove, Pigeon River) Categorical
Latitude Latitude coordinate of the site Numeric
Longitude Longitude coordinate of the site Numeric
DateTime Timestamp of data observation Date (YYYY-MM-DD)
Total Daily Precipitation Daily total precipitation measured at the site Numeric
Mean Discharge Mean stream discharge over the day Numeric (cubic meters/sec)
Mean Gage Height Mean water level height above the gauge reference point Numeric (feet)

0.4 Data Wrangling

To set up the analysis for the selected sites, we retrieved data from the USGS using the “dataRetrieval” package. We began by loading the necessary libraries, including “dataRetrieval”, “dplyr”, and “tidyr” and other relevant packages for analysis. We then defined the site numbers and parameters of interest. The site numbers are: Sugar Grove (03479000), French Broad River in Asheville (03451500), and Pigeon River near Canton (03456500).

We focused on the parameters stream discharge (00060), precipitation (00045), and gage height (00065). The statistical codes include the daily mean (00003) for discharge and gage height, and the daily sum (00006) for precipitation.

To streamline the data retrieval process, we created a custom function using “readNWISdata” from the “dataRetrieval” package to fetch data for the specified sites, parameters, and stat codes. The function should then clean the data by removing unnecessary “cd” columns, which indicate whether the data is official or provisional, as these columns do not significantly affect the analysis. We then renamed the remaining columns to include the parameter and statistical codes, ensuring clarity and consistency across datasets.

We applied the function to retrieve and clean data for each site: Sugar Grove, French Broad River (Asheville), and Pigeon River. This approach ensures uniformity and consistency in the data preparation process. After cleaning, we combined the datasets into a single, cohesive dataset, confirming that all columns are consistently labeled. This unified dataset facilities for effective cross-site comparisons.

Retrieving Data

Retrieving data for Sugar Grove Gage from USGS using USGS’s “dataRetrieval” package. We found its siteNumber is 03479000 from “https://waterdata.usgs.gov/monitoring-location/03479000/#parameterCode=00060&period=P7D&showMedian=false”.

Retrieving data for Asheville Gage from USGS using USGS’s “dataRetrieval” package. We found its siteNumber is 03451500 from “https://waterdata.usgs.gov/monitoring-location/03451500/#parameterCode=00060&period=P7D&showMedian=false”.

Retrieving data for Pigeon River Gage from USGS using USGS’s “dataRetrieval” package. We found its siteNumber is 03456500 from “https://waterdata.usgs.gov/monitoring-location/03456500/#parameterCode=00060&period=P7D&showMedian=false”.

0.5 Daily Time Series Analysis: Daily and Historical Trend Patterns

How unusual was Helene?

First let’s create separate layers from our data for 2024 and then Helene as an event within 2024 a 7 day span from 9/22/24 to 9/29/24.

Investigating Helene: Let’s see how the week of Helene compares to the rest of 2024 within our parameters

Daily Precipitation Trends (Inches) for 2024

Figure 1: Daily Precipitation Trends (Inches) for 2024

This figure illustrates the comparison of daily precipitation trends (in inches) across Sugar Grove, Asheville, and Pigeon River for 2024. During the week of Hurricane Helene, we observed a peak of 5 to 6 inches in all three counties.

It is important to note that Asheville is missing daily precipitation data from April 2024. We do not have an explanation for the lapse in data collection from this USGS stream gage.

Daily Discharge Trends (cfs) for 2024

Figure 2: Daily Discharge Trends (cfs) for 2024

Daily Discharge Trends (cfs) for 2024

Figure 3: Daily Discharge Trends (cfs) for 2024

As shown in Figure 1, this figure compares the daily discharge trends (in cubic feet) across Sugar Grove, Asheville, and Pigeon River for 2024. During the week of Hurricane Helene, we observed peak discharges of 4,000 cubic feet for Sugar Grove, 20,000 cubic feet for Asheville, and 6,000 cubic feet for Pigeon River. It is important to note that Asheville is missing daily discharge data from October 2024. We do not have an explanation for the lapse in data collection from this USGS stream gage. This missing data could potentially impact our Hazen calculations later in the study

Daily Gage Height Trends (cfs) for 2024

Figure 4: Daily Gage Height Trends (cfs) for 2024

As shown in Figure 1 and 2, this figure compares the daily gage height trends (in feet) across Sugar Grove, Asheville, and Pigeon River for 2024. During the week of Hurricane Helene, we observed peak gage heights of 10 feet for Sugar Grove, 15 feet for Asheville, and 8 feet for Pigeon River.

Clearly, Helene was the most significant event for these parameters in 2024.

How do these maximums compare to all historical data we have access to?

Let’s now zoom out to compare these dates with all historical data. This will help us understand how the 2024 data stands in the context of past events.

Historical Daily Precipitation (Inches)

Figure 5: Historical Daily Precipitation (Inches)

This plot depicts the historical trend of daily precipitation (in inches) for Sugar Grove, Asheville, and Pigeon River. The highest peak, recorded during Hurricane Helene, highlights the substantial impact of this climate event on these regions. Red dots mark the week of Hurricane Helene within the historical data, further emphasizing its significance compared to long-term trends.

Historical Daily Discharge (cfs)

Figure 6: Historical Daily Discharge (cfs)

The plot displays the historical daily discharge, measured in cubic feet per second (cfs), for three locations: Sugar Grove, Asheville, and Pigeon River. Asheville exhibits the most significant discharge fluctuations, with peaks exceeding 60,000 cfs. In contrast, Sugar Grove shows a more stable pattern. However, both regions were significantly impacted on August 11th, 1940, by the decade’s most deadly Atlantic storm, a Category 2 hurricane (Roth, 2024). Pigeon River generally has lower discharge and less pronounced peaks, but it experienced the highest discharge this year due to Hurricane Helene.

Historical Mean Daily Gage Height (ft)

Figure 7: Historical Mean Daily Gage Height (ft)

Historical Mean Daily Gage Height (ft)

Figure 8: Historical Mean Daily Gage Height (ft)

The plot visualizes the historical mean daily gage height for three locations: Sugar Grove, Asheville, and Pigeon River. Sugar Grove indicates some large fluctuations in gage heigh, though overall the pattern seems comparably stable. Asheville exhibits substantial fluctuations with notable peaks exceeding 15 feet. Pigeon River generally has lower gage heights overall and less pronounced peaks than Asheville. Both Asheville and Pigeon River recorded their highest gage heights during Hurricane Helene.

These graphs reveal that Helene was responsible for some outlier behavior, but it doesn’t account for seasonality and other trends

0.6 Seven-Day Time Series Analysis: Seasonality Patterns

Let’s decompose these trends to better understand them. It’s worth noting that this analysis will focus on daily discharge data, as it spans from 1940 to the present. In contrast, precipitation and gage height data are only available from the late 1990s onward.

Decomposition of 7-Day Moving Average of Daily Discharge for Sugar Grove

Figure 9: Decomposition of 7-Day Moving Average of Daily Discharge for Sugar Grove

This plot shows the decomposition of the 7-day moving average discharge data for Sugar Grove. The top plot represents the seasonal component for discharge, representing periodic fluctuations in discharge based on seasonal variations at Sugar Grove. The middle plot show the underlying trend of discharge over time, reflecting the long-term changes in discharge patterns. The bottom plot displays the irregular fluctuations that are not explained by seasonal or trend components, highlighting the significance of rare high-immpact occurrances like Hurricanes.

Decomposition of 7-Day Moving Average of Daily Discharge for Asheville

Figure 10: Decomposition of 7-Day Moving Average of Daily Discharge for Asheville

Decomposition of 7-Day Moving Average of Daily Discharge for Pigeon River

Figure 11: Decomposition of 7-Day Moving Average of Daily Discharge for Pigeon River

The graphs reveal important patterns, highlighting 2024 as a potential statistical outlier. However, they do not quantify the extent to which Hurricane Helene deviates from historical norms. To assess this with statistical rigor, we aim to determine how unusual Helene was as an event.

To achieve this, we will use the Hazen Method to calculate the recurrence intervals for Hurricane Helene within the analyzed parameters. This method will help contextualize the storm’s magnitude and frequency relative to historical data.

Our first step will be calculating recurrence intervals for 7-day precipitation events. Based on initial analysis, it is reasonable to begin by assuming that Hurricane Helene represents the 7-day averaged maximum for 2024.

Helene Recurrence Intervals: Hazen Analysis

Table 2: Sugar Grove Precipitation Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 1.56 1 5.56 18

The Hazen number of 18 indicates that the maximum 7-day average precipitation event in 2024 was an 18-year event, meaning it had about a 5.5% chance of occurring in any given year (Hazen interval: 18 years, probability: 5.5%).

Table 3: Asheville Precipitation Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 1.76 1 1.85 54

The Hazen number reveals that the 7-day average discharge observed in Asheville in 2024 qualifies as a 54-year event. In other words, such an event is expected to occur approximately once every 54 years, equating to a 2% probability of happening in any given year. (Hazen interval: 54 years, probability: 2%)

This provides a statistical metric to quantify the observed outliers.

Next, we will calculate the recurrence intervals for 2024 at our last selected site.

Table 4: Pigeon River Precipitation Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 1.98 1 1.92 52

At Pigeon River, Hurricane Helene was a 52-year event in terms of total precipitation. This indicates that Helene’s precipitation was less unusual at Sugar Grove than it was in Asheville or Pigeon River during the maximum 7-day event. (Hazen interval: 52 years, probability: ~2%)

Let’s now automate this function for all the rest of parameters and run our sites.

Table 5: Sugar Grove Discharge Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 1138 28 32.35 3.09

The table shows that for Sugar Grove in 2024, the maximum event had a Hazen value of 3.09. This indicates that the event had a return period of approximately 3.09 years, meaning an event of this magnitude or greater is expected to occur, on average, once every 3.09 years. In other words, the event was not particularly unusual or extreme for Sugar Grove (Hazen interval: 3.09 years, probability: ~32%).

Table 6: Asheville Discharge Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 8778.57 52 41.2 2.43

The table shows that for Asheville in 2024, the maximum event had a Hazen value of 2.43. This indicates that the event had a return period of approximately 2.43 years, meaning an event of this magnitude or greater is expected to occur, on average, once every 2.43 years. In other words, the event was not particularly unusual or extreme for Asheville (Hazen interval: 2.43 years, probability: ~40%).

Table 7: Pigeon River Discharge Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 1815.71 2 2.11 47.33

The table shows that for Pigeon River in 2024, the maximum event had a Hazen value of 47.33. This indicates that the event had a return period of approximately 47.33 years, meaning an event of this magnitude or greater is expected to occur, on average, once every 47.33 years. In other words, the event was particularly unusual or extreme for Pigeon River compared to Asheville and Sugar Grove (Hazen interval: 47.33 years, probability: ~2%).

Table 8: Sugar Grove Gage Height Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 3.54 16 51.67 1.94

The table shows that for Sugar Grove in 2024, the maximum event had a Hazen value of 1.94. This indicates that the event had a return period of approximately 1.94 years, meaning an event of this magnitude or greater is expected to occur, on average, once every 1.94 years. In other words, the event was not particularly unusual or extreme for Pigeon River (Hazen interval: 1.94 years, probability: ~52%).

Table 9: Asheville Gage Height Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 11.25 1 1.28 78

The table shows that for Asheville in 2024, the maximum event had a Hazen value of 78. This indicates that the event had a return period of approximately 78 years, meaning an event of this magnitude or greater is expected to occur, on average, once every 78 years. In other words, the event was particularly unusual or extreme for Asheville (Hazen interval: 78 years, probability: ~1.3%).

Table 10: Pigeon River Gage Height Hazen Values for 2024
Year Max Event Rank Fa Value Hazen Value
2024 3.55 2 5 20

The table shows that for Pigeon River in 2024, the maximum event had a Hazen value of 20. This indicates that the event had a return period of approximately 20 years, meaning an event of this magnitude or greater is expected to occur, on average, once every 20 years. In other words, the event was particularly unusual or extreme for Pigeon River (Hazen interval: 20 years, probability: ~5%).

Map of Gage Sites

To visually illustrate how unusual Hurricane Helene was at each site, we will create a map that represents the magnitude of the Hazen interval.

The size of the dot on the map will correspond to the rarity of the event: the larger the dot, the less likely the event was to occur.This visualization will help convey the statistical significance of Helene’s impact across the selected sites.

Figure 12: Hazen Value (Size)

What we found was very interesting!

Now that we have an easy function, we will download more sites in Western North Carolina that have precipitation, discharge, and gage height available. This will be a helpful visual in understanding the specific impacts of Helene in Sugar Grove, Asheville,and Pigeon River compared to the rest of the state.

After selecting the additional sites, they were incorporated into an array to enhance the map.

The map now displays the magnitude of each Hazen interval, representing how unusual the 7-day Hurricane Helene event was at each site. The size of the dot corresponds to the rarity of the event: larger dots indicate more unusual 7-day averages for the given parameter. Although the number of sites with all three parameters is limited, further analysis could enable the creation of increasingly granular data for each site, improving the resolution of our insights.

Why this analysis matters

This map serves as a representation of relative intensity compared to historical norms. While the size of the dots does not directly indicate the absolute intensity of precipitation or flooding, it reflects the magnitude of these parameters relative to each area’s historical averages.

Flood management infrastructure is typically designed to handle storms with a return period of 50 or 75 years. Therefore, understanding the relative magnitude compared to historical norms is crucial for assessing how well the infrastructure performed under the conditions of Hurricane Helene—or how severely it was overwhelmed.

Although the number of sites with all three parameters is limited, the analysis reveals that Helene’s relative intensity to historical norms was greatest along a path through central Western North Carolina, particularly over Asheville. Based on this, we can reasonably infer that Asheville’s infrastructure faced greater inundation compared to areas in northeastern North Carolina or other sites included in this study.

0.7 Summary and Conclusions

The analysis of Hurricane Helene in 2024 reveals it was a significant event across multiple parameters, particularly precipitation, discharge, and gage height. Comparing daily trends in precipitation and discharge across Sugar Grove, Asheville, and Pigeon River, Helene caused peak values of up to 6 inches of rainfall, 60,000 cfs of discharge, and 15 feet of gage height in Asheville. The temporal and spatial differences in rainfall intensity are highlighted, showing that Asheville saw the most extreme impact. The data from these three locations provide a comparative view of how the storm impacted different regions, with Asheville seeing the most substantial effects in terms of both rainfall and gage height.

When we extended this analysis to historical data, we found that Helene’s precipitation and gage height were exceptionally high. The Hazen Method applied to historical data indicates that the storm was a rare event in terms of precipitation for Asheville and Pigeon River, marking a 52-78 year event. In contrast, the discharge levels were less unusual for Sugar Grove and Asheville, with Hazen values indicating these events were expected every 2-3 years. This suggests that while Helene was significant, particularly in precipitation and gage height, the discharge levels were relatively typical for these regions.

To further contextualize the significance of Helene, we applied the Hazen Method to calculate recurrence intervals, revealing that the storm was a statistical outlier in terms of precipitation for Pigeon River and Asheville, but more typical in Sugar Grove. The Hazen values highlighted the extreme nature of Helene’s impact, particularly in Asheville, where the maximum precipitation event in 2024 was a 54-year occurrence. A visual map of Hazen values reinforced these findings, showing that the storm’s relative intensity was greatest in central Western North Carolina, especially Asheville, suggesting that the region’s infrastructure likely faced more significant challenges in managing the storm’s impacts compared to other areas in the study. This analysis provides valuable insights into the magnitude of the event and its implications for flood management and infrastructure in the region.

In conclusion, this analysis of Hurricane Helene’s impact across Sugar Grove, Asheville, and Pigeon River revealed distinct rainfall and flooding patterns. Asheville experienced the most severe effects, with extreme precipitation and flood levels, while Sugar Grove saw more typical discharge levels. The storm was a rare event for Asheville and Pigeon River, but more typical for Sugar Grove. While we can analyze data and predict weather trends, it is important to note and remember the people who have been and continue to be impacted and affected by weather events. This analysis highlights the need for region-specific flood management strategies, considering historical data and projected weather patterns, and the unique needs of local communities.

0.8 References

Cooper, R. (2024). Hurricane Helene Recovery Recommendations Preliminary Damage and Needs Assessment. https://www.osbm.nc.gov/hurricane-helene-dna/open

North Carolina Office of State Budget and Management (OSBM, 2024). Hurricane Helene DNA.” https://www.osbm.nc.gov/hurricane-helene-dna/open (accessed December 09,2024)

Roth, D. (2024). Southeast Hurricane - August 10-18, 1940. Noaa.gov. https://www.wpc.ncep.noaa.gov/tropical/rain/sehrcn1940.html

Thiem, H., & Lindsey, R. (2024, November 7). Hurricane Helene’s extreme rainfall and catastrophic inland flooding. NOAA Climate.gov. https://www.climate.gov/news-features/event-tracker/hurricane-helenes-extreme- rainfall-and-catastrophic-inland-flooding

Please Note: Portions of code created in cooperation with “R Wizard” AI on OpenAI’s ChatGPT platform. Limited prompts available upon request.