Red Flag Days in California - A Hasty Climatology

Introduction

With all the California wildfires in the news, and questions over whether these fires can be attributed to climate change, human activity (e.g. arson), or forestry management, I decided to take a look at the trends in the number of Red Flag days over the past few decades. "Red Flag" conditions are weather conditions that when combined with critical fuels (e.g. dead trees, grass build-up, etc.) and an ignition source (e.g. lightning, an ill-placed cigarette butt, etc.), will foster rapid fire growth. In general, the two main things you need for rapid fire growth are low humidity and high winds. The criteria for the National Weather Service to issue a Red Flag Warning will vary depending where in the country you are, and the criteria may be eased (either direction) a bit depending on fuels. For example, a local weather forecast office (WFO) may issue a warning below criteria if fuels are really primed (such as during a severe, long-term drought). Conversely, a WFO may be more conservative on issuance if fuels are very wet from say recent heavy rainfall. Then on top of that, things like potential for dry lightning (i.e. thunderstorms that produce lightning but no rainfall, which are common in the West) have to be considered. After all of that, looking around the internet, it seems like the roughly agreed upon definition amongst the National Weather Service WFOs in California is humidity of less than 15% and wind speeds of more than 25 MPH. Keep in mind that warm and windy conditions are just one part of a complicated equation!

Literature Review

Interestingly, there doesn't seem to be a ton of academic research on this topic. I searched the journals of the American Meteorological Society, and in particular the Journal of Climate. Most of the recent papers on the topic included one on dry lightning climatology, short-term climatology (i.e. intraseasonal variability), and droughts and tree mortality. I'm sure there is much more out there, and I humbly ask if you the reader knows of any papers on this topic, to send them my way!

Buyer Beware

I am by no means an expert on California climate and wildfires, nor would I consider myself a climatologist. I merely decided to look at this data because a lot of hay has been made over this issue in the news and among the talking head crowd. I will also admit that this isn't the most robust analysis in the world because of the fuel and ignition issues mentioned above, as well as the fact we are looking at point observations. California has notoriously complex terrain and micro-climates, so an observation site even just 10 miles away from a wildfire may not be particularly representative.

Methodology

The computation of Red Flag days essentially involved four steps. And for more details, my code is available on the website.

1. Archived METAR data is downloaded from Iowa State University's repository, after which each observation is determined to either be Red Flag ("True") or not ("False"). I used relative humidity of less than 15% and wind speeds of greater than 25 MPH, but these settings are user editable.
2. The date and year are extracted from the valid time, ignoring the time of day. We only care about the number of days. This is one short-coming I will admit: no distinction is made between a day where one red flag observation occurs for an hour, or a day where it occurs for several hours. This would make the code more complicated, but probably could be done.
3. The individual dates of red flag conditions occurring are determined.
4. We group those dates by year, and determine how many dates red flag conditions occurred within a given year.

Results

The graphs below were posted here because they are all in locations that are in or near the mountains of California, and had at least 30 years of data, and had at least 10 years where there was at least one Red Flag day.

Alturas, Modoc County (Northern California - Interior)

Campo, San Diego County (Southern California - Interior)

Mammoth Lakes, Mono County (Sierra Nevada)

Sandberg, Los Angeles County (Southern California - Transverse Ranges)

See the complete collection here!

Commentary

Three of the four sites above (Campo, Mammoth Lakes, and Sandberg) show an uptick in, or already a large number of existing Red Flag days. Alturas on the otherhand shows little trend, but perhaps because it is in the far northwestern corner of California where the Santa Ana winds tend to be less of a factor. Looking at the other plots, we can also see noticeable upticks in Barstow, Edwards Air Force Base, Needles, China Lake, El Centro, Twentynine Palms, and Palmdale. One thing all these sites have in common is they are interior locations in Southern California. The trends are less pronounced from what I can tell in Northern California.

Despite the shortcomings of this analysis I have mentioned, it does appear there is at least a little bit of legitimacy to the claim that California is seeing more days conducive for rapid fire growth over the last 30 years. This seems to be particularly true in interior Southern California, though as I said, less so in Northern California (with Mammoth Lakes in the high Sierra being a noteworthy exception).

So whether it is forest management, human activity, or weather, all three of these are very important when it comes to long-term fire prevention strategies. I am not going to comment on any policy implications of what this all means, as it is a vastly complex issue, and I am only an operational meteorologist! Not an ecologist, economist, or any of the other experts that should have a seat at the table in such discussions.