Climate Change Analysis in the First State

Esteemed members of the Wilmington City Council, this analysis has been conducted to understand climate trends within your humble city better. Wilmington is a great city to consider for climate analysis because of how detailed and readily available data is. Daily temperature data of highs and lows dating back to the 1940s gives us a unique opportunity to assess what has changed in that time frame. This analysis looks at temperatures, highs, and lows from 1950 to April 17th, 2022. The National Oceanic and Atmospheric Administration (NOAA) has collected the data. By requesting this link, the data was provided in a format our analysis could use.

https://www.ncdc.noaa.gov/cdo-web/datasets/GHCND/stations/GHCND:USW00013781/detail

This portion of the code pulls a subset within this extensive and rich dataset, precisely the date field and the high and low temperatures. The date is also being transformed from a character field to a date field so modeling and visuals can be constructed.

rawData_df <- read.csv("2951994.csv")

cleanerDF <- rawData_df[, c('DATE','TMAX','TMIN','PRCP')]

cleanerDF$DATE <- as.Date(cleanerDF$DATE, "%m/%d/%Y")


This first visual shows the high and low temperatures from the dataset and how they group. Temperatures seem consistent, but there are a few days where the highs were in the 60s and the lows were in the teens, and a few data points where the highs were in the high 60s and low 70s, but the low temperatures for those days were in the 30s. These temperature call-outs are worth mentioning because Wilmington is near Delaware Bay and is not a far drive to some exquisite beaches. Coastal and adjacent coastal cities typically have more moderate temperatures; these outliers are apparent exceptions to that expected climate.


Link to video presentation - https://www.youtube.com/watch?v=oJLBwCz3Uos

Temperatures over 72 Years

These scatterplots show daily highs and lows over the 70-year timeframe. The blue line of best fit shows a subtle upward slope indicating that the daily highs and lows are slowly trending upward, indicating temperatures are increasing in Wilmington. A Linear regression model was run for each measurement measured against time to understand these charts and their meanings better. For the first regression model, the slope is 0.00007177 with a y-intercept regarding the time dimension of 63.65. The hill is slightly less than .0001 degrees Fahrenheit but is close to the 2 degrees Celsius per century temperature increase climate scientists worry about. The second regression model was done with the daily low temperatures, and the slope for that model was higher, with .0001074 and a y-intercept of 44.69. These models had low multiple R-Squared values of .000865 for the first model and .002303 for the second, indicating that the date variable alone does not explain the temperature variability best.

regressionModelHiTemp <- lm(TMAX~ DATE, data = cleanerDF)

summary(regressionModelHiTemp)
## 
## Call:
## lm(formula = TMAX ~ DATE, data = cleanerDF)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -60.285 -14.996   1.521  16.178  38.750 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 6.365e+01  1.429e-01  445.45  < 2e-16 ***
## DATE        7.177e-05  1.483e-05    4.84  1.3e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 18.37 on 26403 degrees of freedom
## Multiple R-squared:  0.0008865,  Adjusted R-squared:  0.0008486 
## F-statistic: 23.43 on 1 and 26403 DF,  p-value: 1.305e-06
regressionModelLowTemp <- lm(TMIN~ DATE, data = cleanerDF)

summary(regressionModelLowTemp)
## 
## Call:
## lm(formula = TMIN ~ DATE, data = cleanerDF)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -59.279 -13.574  -0.408  14.992  35.682 
## 
## Coefficients:
##              Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 4.469e+01  1.326e-01 337.017  < 2e-16 ***
## DATE        1.074e-04  1.376e-05   7.807 6.07e-15 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 17.04 on 26403 degrees of freedom
## Multiple R-squared:  0.002303,   Adjusted R-squared:  0.002265 
## F-statistic: 60.95 on 1 and 26403 DF,  p-value: 6.065e-15

Temperatures in August 2019

This scatterplot showed daily high temperatures for August 2019. This is a more recent timeframe, and easier to pinpoint current policies impacted by climate change. The scatterplot below it shows temperatures for the same month but in 1955. Some callouts are that there were hotter days in 1955, but the temperature trended downward as the month progressed, and the temperatures in 2019 remained higher for longer. More water is needed to mix concrete to support infrastructure for road work and construction projects. Wilmington may need to consider scheduling construction projects in tighter timeframes to offset the excess summer heat and the restrictive cold winters.

Temperatures in August 1955

This visual is a scatterplot of daily high temperatures for April 16th from 1950 to 2022. This date is significant for my birthday, a spring data point for this analysis. This plot illustrates the same trends above, but the outliers are interesting. One year in the early 2000s was almost 90 degrees, and in the early 1960s, there was a data point in the low 40s.

The temperature variability and the interquartile range will help show how temperature has trended in Wilmington over the past decades. This first plot shows the month of August in the 1950s and every subsequent ten years after in our dataset. The box and whisker plot illustrates the range of the data for each year. This range has the max and min data points and the interquartile range boxes. The bottom part of the box is the lower quartile, the line above represents the median value for that year, and the top part is the upper quartile for that year. The dot inside the box represents the median value. The dots seen above or below the whisker lines show the possible outliers for that year.

The first box and whisker plot shows August’s daily high temperatures over the decades. The 1990 August data point had the most outliers with daily highs below 70, which is quite cool weather for the month. There was an over-daily high-temperature drop between 1990 and 2000, but it trended upward in 2010 and 2020. The daily min temperature box plot reflects this trend as well. What stands out with this chart is that there was a spike in 1970, but the majority of data points below that data point regarding the box’s position. There was a temperature dip in 1990, but the boxes trended upward in 2000, 2010, and 2020. These daily low temperatures are getting higher, which was discovered in the earlier steps of the analysis.

The daily temperatures for the highs and lows for August are shown in the table below. The Delta column is the daily difference between high and low-temperature variance. As this number increases, there are more days where the daily high and low temperatures vary. This means higher energy costs from a resource planning perspective because more people are running air conditioning. Based on the box plots findings above, people may have days where they had run their furnace in August. These temperature fluctuations will put a tremendous strain on energy providers for Wilmington and cost its residents more in utilities. Adopting more energy-efficient practices while pursuing renewable initiatives will help offset the long-run costs associated with these changes in the climate.

Year High Temp Low Temp Delta
1950 25.09892 26.698920 7.86237
1960 22.97849 15.746240 18.29032
1970 12.15914 14.795700 10.79785
1980 31.79570 17.522580 25.77849
1990 39.64731 9.049462 40.03656
2000 31.56129 45.722580 31.46452
2010 36.80645 26.513980 47.51183
2020 20.20645 17.845160 14.38065

Factoring precipitation into this analysis is also beneficial, especially if droughts are a remote concern; this could mean increased water utility costs, drought-resistant landscape planning, and adaptation of different community garden horticulture techniques.