Dark Winds, Omens & Reality: Testing Joan Didion’s Theories on the Santa Ana Winds
Authors: Jessica Ye (Project Lead), Brycen Izu, Ian Kim, Ella Hinkle, Jaiden Ha
Introduction
In her 1969 essay The Santa Anas, the late American writer Joan Didion paints a foreboding picture of the seasonal, desert-born winds that sweep through Southern California — winds that have once again fueled catastrophic wildfires. In early January of this year, the destructive Palisades Fire, driven by the dry and relentless Santa Ana gusts, burned for 24 days through Pacific Palisades, Topanga, and Malibu. However, Didion argues that the winds carry more than just the threat of wildfires.
Beyond just a meteorological event, Didion claims that the winds are a bad omen, a force that creates unease, frays tempers, and even incites violence. “To live with the Santa Ana,” Didion writes, “is to accept, consciously or unconsciously, a deeply mechanistic view of human behavior.”
It’s a compelling claim, one echoed in local lore and pop culture, but does data support the idea that the Santa Ana winds influence crime, accidents, or even our collective mood? With access to modern datasets — crime reports, emergency response logs, and the Santa Ana Winds Regional Index — we can explore whether the winds truly coincide with surges in unrest or if this is simply a case of confirmation bias.
The Santa Ana Winds Regional Index (SAWRI), spanning seven decades, provides a comprehensive record of Santa Ana wind activity across Southern California. In this article, we will use this dataset to explore Didion’s claims and assess whether there is a data-driven connection between the seasonal winds and the behavioral effects on the people of Los Angeles County.
Do the Santa Ana Winds Cause an Uptick in Crimes?
In her article, Joan Didion described a period in 1957 when the Santa Ana winds were unusually strong and persisted for exactly fourteen days. During this time, she recounted a series of violent crimes that occurred during this period, including the brutal murder of a 22-year-old South Gate divorcee, who was brutally killed and thrown from a moving car.
To investigate the potential link between the Santa Ana winds and an increase in criminal activity, we analyzed two datasets: one containing records of reported crime incidents in Los Angeles from 2010 through September 2017 and another documenting arrests made in LA between 2010 and 2019. The two graphs shown below illustrate the relationship between the SAWRI on a given day and the number of crimes committed or arrests made on that day.
Neither scatterplot shows a clear correlation between crime or arrest counts and the SAWRI, as the data appears randomly distributed. While Didion’s account highlights specific violent incidents during the Santa Ana winds, these appear to be anecdotal rather than indicative of a broader pattern. As there is no evident statistical relationship to support the claim that the wind intensity is correlated with crime rates, it suggests that other factors may be influencing the specific violent incidents Didion describes.
Additionally, it is important to note that the datasets analyzed contained data primarily from 2010–2017/2019, whereas Joan Didion refers to events that occurred in 1957. Changes in social conditions, law enforcement practices, and environmental factors over time may limit the applicability of these findings to her claims.
Santa Ana Winds vs Domestic Violence
In the article, Didion details several cases of family annihilations and acts of domestic violence during the Santa Ana season. Cross referencing the SAWRI database and the California Department of Justice’s domestic violence calls over the past twenty years, we see a discrepancy.
The visualization presents two key trends: the average SAWRI by month for different decades (plotted in various colors) and the normalized domestic violence call rates aggregated by months as well (dashed red line).
While the SAWRI data follows a consistent seasonal pattern, peaking from October through February and tapering off almost entirely during the summer months, the normalized domestic violence call rates exhibit a different temporal pattern. While there is a slight increase in reported DV cases during the peak wind season, most of the calls occur during the summer, a period when the SAWRI is effectively zero.
This contradicts the direct causation implied in Didion’s essay, where she suggests that the Santa Ana winds exacerbate tensions and contribute to heightened violence in homes. Instead, the data points towards a more complex relationship between domestic violence and seasonal factors.
The summer spike in calls could be influenced by other external variables such as increased time spent at home due to school closures, rising temperatures, or economic stressors rather than the presence of the winds alone. Meanwhile, the milder increase in domestic violence calls during the wind season may indicate some correlation but does not support a deterministic effect of the winds on violent behavior.
Police Shootings in Windy Weather
The trend of violence detailed in Didion’s article raises the question of whether authoritative confrontations with the public increase during Santa Ana wind months. In theory, Didion’s claims that the winds contribute to heightened tension or erratic behavior could lead to more frequent police encounters, escalating into shootings.
To investigate a possible correlation between the SAWRI and the number of police shootings, we analyzed cities affected by the seasonal winds from a dataset that documented police shootings across multiple different cities in multiple different states.
This double line plot explores whether the frequency of police shootings fluctuated in relation to the SAWRI. The graph presents the quarterly average Santa Ana Winds Regional Index and number of police shootings from January 2015 to December 2018 in cities directly affected by the Santa Ana winds, including Santa Ana, Los Angeles, San Diego, etc. Each quarter consists of three months, where Q1 is January, February, and March; Q2 is April, May, and June, and so on.
Contrary to our hypothesis, the data appears to be inversely related. Notably, shootings tend to decrease when the index is high, such as in Q4 of 2017 and 2018, and increase when the index is low, as seen in Q3 of 2016 and 2018.
To quantify this correlation, we calculated the Pearson correlation coefficient, which measures the strength and direction of a linear relationship between two variables. The coefficient for this data is -0.270 with a p-value of 0.313, indicating a weak negative correlation between SAWRI and number of police shootings. While this suggests some degree of inverse relationship, the weak correlation and lack of statistical significance point towards other factors that likely play a more dominant role.
A more refined approach would involve further research into cities affected by Santa Ana winds, grouping them based on the severity of their exposure, and re-examining the relationship between police shootings and the SAWRI in the most impacted areas.
Deaths from the Santa Ana Winds
Multiple instances of death in Los Angeles are ostensibly linked to the Santa Ana winds in Didion’s article, with her recounting how many deadly traffic accidents and violent crimes occurred during “the longest single Santa Ana period in recent years” at the time of writing her piece in 1969. Consequently, we believe examining the relationship between the frequency of deaths in LA County and the SAWRI at the time they occurred is relevant for gauging the truth of her storytelling.
To conduct this, we obtained monthly counts of deaths in LA County from 2004 to 2018 from the CDC’s Multiple Cause of Death database. These records, which include underlying causes of death, allow us to categorize deaths by cause (e.g., circulatory-system diseases, external causes of mortality). We then used simple linear regression models to analyze how monthly death counts correlated with the corresponding average SAWRI, helping us assess whether Santa Ana winds influence different causes of death.
This plot visualizes the regression models for each of LA County’s top five causes of death. Notably, monthly SAWRI holds a significant positive relationship with deaths from the leading cause of death, diseases of the circulatory system. Specifically, monthly deaths from circulatory system diseases increase by an average of 112.51 for every one unit increase to the SAWRI of that month — a correlation indicated to be statistically significant.
In contrast, external causes of mortality, including suicides, accidents, and violent crime, do not show a positive relationship with monthly SAWRI. In fact, they exhibit a slight negative correlation where deaths by external cause decreased by an average of 4.932 for every one unit increase in the SAWRI of that month. This negative correlation was also marked to be significant using a statistical test, albeit only barely.
Closer analysis of specifically suicide data unveils statistically insignificant relationships between monthly suicides in LA County and SAWRI, regardless of the type of suicide.
Consequently, while there is evidence to suggest that deaths from circulatory and respiratory diseases diseases increase during the Santa Ana wind season, Didion’s story about how suicides and other deaths of violent nature escalate with the Santa Ana winds is not supported with our analysis.
Feeling the Santa Ana Winds
Also in her article, Didion writes that despite not hearing that the Santa Anas are due, people “know it because [they] feel it.” She goes on to list possible phenomena mentioned previously and the devastating LA fires from throughout the years as reasons for people “feeling” the Santa Ana winds coming. By investigating the search patterns of LA residents during months with high SAWRI, we can see if people are subconsciously aware of the relationship between high winds and their fear and knowledge of wildfires. Fire-related terms were chosen since fires are closely associated with the Santa Ana winds in media and society. The following graphs map the average Google interest (or popularity) for the term compared with the average SAWRI index per month.
The first term, “Fire Safety”, appears to spike in October, which is the first month Santa Ana winds begin to appear after they die down after March. Besides this spike from August to November, the average interest stays below 32 for the rest of the year.
For the term “Arson,” interest peaks in January, one of the highest months of SAWRI. Similarly, there is a peak in the search interest for “Arson” during October, which, as said above, marks the first month of seeing Santa Ana winds. December also sees a spike in interest, but overall, the variation only ranges from a little above five to ten, with the majority of the months having interest around 7.
When looking at the term “Wildfire,” its interest also peaks during October. Regardless, search interest remains high from June to December, and there is only a total change of four in interest from the lowest to the highest month.
In all three graphs, there is a spike in fire-related topics during October, the first month with a non-zero Santa Ana Wind Regional Index. This means that the first time the Santa Ana winds could or do appear in the second half of the year is when people subconsciously search for fire-related information, suggesting that maybe Didion was right about people “feeling” the (initial) change. However, the search trends don’t align for the majority of the terms otherwise and don’t follow the uptick in Santa Ana winds. There is no significant change in the Google search trend for “Arson” or “Wildfire,” and there could be confounding factors, like the number of fires in a given month. This uncertainty leaves us with no conclusion, just a hint that the first time the Santa Ana winds appear leads to a change in awareness about fires.
Conclusion
Our analysis finds that Joan Didion’s portrayal of the Santa Ana winds as a harbinger of unrest and violence is more anecdotal than statistically supported. While her evocative descriptions capture the psychological and cultural perception of these seasonal winds, the data does not consistently show a positive correlation between the Santa Ana winds and increased crime, domestic violence, or police shootings. In fact, some trends, like the negative correlation between police shootings and high-risk wind periods, directly contradict her claims.
However, there are some notable findings. The strongest correlation we discovered was between Santa Ana winds and circulatory-related deaths, which suggests that these winds may have some sort of physiological effect on human health. Additionally, Google search patterns display an increased awareness of fire risks during the return of the winds in fall, lending some credence to Didion’s idea that people can “feel” the Santa Anas before it comes.
While Didion’s claims ultimately may not hold under rigorous statistical scrutiny, her essay remains a poetic reflection and a cultural homage to Southern Californian history. The text encapsulates the psychological weight the Santa Ana winds carry, reflecting the region’s vulnerability to fire and myth.
