A Practical Guide to Understanding and Recreating the 1906 San Francisco Earthquake Experience
The 1906 San Francisco earthquake is one of the most well-documented seismic events in American history. If you are looking to understand it deeply, roleplay it, or create educational content around it, there is a specific set of resources and methods that actually work. Most people go about it wrong by relying on entertainment-grade simulations that get the physics and social dynamics completely wrong. Before we get into the practical side, it helps to know what we are dealing with. The earthquake struck at 5:12 AM on April 18, 1906. The magnitude is estimated between 7.8 and 8.3. The shaking lasted roughly 45 seconds to a full minute depending on your location. The fires that followed burned for three days and caused more damage than the ground motion itself. That distinction matters a lot if you are building anything educational or immersive around this event. There is no single official app or downloadable program called "I Survived The San Francisco Earthquake 1906." What exists are several separate resources that people lump together. There is the classic novel by Lauren Tarshis in the "I Survived" series, which is aimed at younger readers. There are various historical simulations, museum exhibits, and educational kits. There are also academic papers, USGS maps, and the famous Charles K. Davis photographic archive showing the destruction.
Here is what most people miss when they try to recreate or study this event. The first thing you need is accurate ground-motion data, not a dramatized movie scene. The USGS has a comprehensive Did You Feel It? database where you can pull actual reported intensity values from different neighborhoods. Those values map to Modified Mercalli Intensity scales, not Richter magnitudes. Using the right scale changes how you model structural damage, and most beginners mess this up by applying uniform destruction across the city. The approach that actually works goes like this. Start with the USGS ShakeMap for a hypothetical repeat of the 1906 event. You can find this at earthquake.usgs.gov. Download the peak ground acceleration and spectral acceleration layers. Then overlay them onto a 1906 San Francisco street map. The Library of Congress has high-resolution scanned maps from that era in their online catalog. Once you have those two layers aligned, you can estimate which structures would have failed and which would have survived based on construction type. Wooden frame buildings performed surprisingly well. The worst damage hit the unreinforced masonry structures along the waterfront and in the business district.
I spent months building a detailed recreation of this for a university project, and the biggest problem I ran into was wind data. Everyone focuses on the earthquake, but the fire spread was driven by wind patterns that day. The actual wind on April 18 was light to moderate from the northwest, but the destruction of the water mains meant fire hydrants were dry. Firefighters tried to create containment gaps by dynamiting buildings, but they could not control which direction the fires jumped. When I tried to simulate this without accounting for the water system failure, my model showed fires burning out within hours. In reality they burned for three days. The fix was simple once I figured it out: remove the water supply constraint entirely from the early phase of the simulation and let wind and fuel load drive the spread instead. If you want a more accessible entry point, the Oakland Museum of California and the California Historical Society both have digitized collections. The California Digital Library holds thousands of photographs, letters, and newspaper clippings. The 1906EBDB project at George Mason University is probably the best structured resource available. It is a fully searchable database of every named person who died in the earthquake and fires, compiled from multiple archival sources. It is not a simulation, but it gives you the human scale that raw seismological data lacks. For people who want an interactive experience, the simplest working method is a combination of Google Earth historical overlays and the Stanford Spatially Explicit Archive. You can walk through a modern satellite view of San Francisco and then pull up 1906 fire insurance maps from the David Rumsey Map Collection. Those maps show building footprints, construction materials, and street names as they existed before the earthquake. Cross-reference a specific block with the Davis photographs showing that same block after the fire, and you get a fairly accurate before-and-after without needing any special software.
Get the Full Details
Here is the counter-intuitive part that most guides skip. The death toll is still disputed, and the commonly cited figure of 3,000 to 4,000 deaths is likely an underestimate. Some historians estimate up to 10,000. The reason is that many victims were never recovered or were burned beyond identification. The official count excluded homeless populations who died in the refugee camps that sprang up in Golden Gate Park and other open spaces. If you are building any kind of educational content, using the standard number without noting this uncertainty will make your work look shallow to anyone who has done real research on the topic. Another thing beginners consistently overlook is the role of the Japan Exchange and Treaty Commission. After the earthquake, Japan sent significant aid including financial assistance and medical supplies. This is often left out of American retellings. Including it matters if you want an accurate picture of the event's broader context. For a hands-on project, the most practical path I found was combining three free tools. QGIS for mapping and spatial analysis, LibreOffice for document assembly, and the Chronicling America newspaper database for period-accurate reporting language. I built a neighborhood-level damage report for the Marina District by taking the 1906 Sanborn fire maps, aligning them in QGIS with modern parcel data, and tagging each building by construction type. The whole process took about six hours once I had the map alignment figured out. The alignment step was the hardest part because the street grid shifted slightly over the decades due to filling projects along the waterfront.
If you are looking for something quicker, the YouTube channel Deep Look and the NOAA educational resources have short documentary-style videos that are accurate enough for casual use. The documentary "The Great Earthquake" from the California Historical Society is also freely available through their website and covers the event with current scholarship. The main pitfall to avoid is treating this as a single event. It is a sequence: the quake, the fires, the quarantine, the refugee crisis, the rebuilding. The rebuilding phase alone lasted over a decade and involved major changes to building codes that affect San Francisco architecture to this day. The 1907 building ordinance required fire-resistant materials for commercial buildings and that is directly responsible for the steel-frame construction boom in the early 1910s. If you stop your coverage at the fires, you are missing half of why this event still matters. There are also physical exhibits you can visit. The 1906 Earthquake Museum at the California Historical Society in San Francisco has rotating displays with primary source material. The Berkeley Seismological Laboratory occasionally runs public demonstrations and has archived lecture recordings online. Neither requires any special access or registration.
The single most useful thing I learned through this process is that the best resource is not a single app or download. It is the habit of cross-referencing at least three independent sources before drawing any conclusion. The 1906 earthquake has been studied for over a century, and nearly every claim you encounter has been debated in academic literature. The versions that survive scrutiny are the ones you can trace back to primary archival documents rather than secondary summaries.
