A Natural Experiment in Sword Fern Mortality

https://pshannon.net/splitSite2025

We have been tracking, and trying to understand dramatic and novel mortality of sword ferns in our old-growth forest’s understory. Only sword ferns (Polystichum munitum) are affected.

The full twelve year project is documented here.

Two Univeristy of Washington students spent the summer surveying throughout the Puget Lowlands of Washington state, establishing that the sword fern blight is intermittently present throughout that region. This project is reported here.

At Seward Park, we have observed the “split site” (see map tab) informally for two years. This summer (2025) we approached it systematically, by

• Mapping and classifying each of the 39 ferns.
• Analyzing the soil in each half

The site provides a natural experiment in that the two halves have

• apparently identical abiotic conditions,
• are equal in sword fern density and count
• are separated only by one meter horizontally
• have the same slope and aspect
• have no obvious canopy differences (though month-by-month insolation data have not been collected)
• are separated only by a fallen Douglas-fir log which, in the site, has a ground clearance averaging three inches (need to verify this!) allowing easy passage air, insects, small mammals and rainwater.

Nonetheless, and for several years, sword mortality is dramatically different between the two halves. The remnant dead fern crowns (dark grey markers on the map) indicate that mortality struck here three-to-five years ago. This summer (I write in September 2025) a few of the “healthy side” ferns show their first signs of blight.

On the affected half-site, the survival of 3 of 21 ferns (14%) fits our informal sense that even in heavily affected pure stands, and perhaps due to genetic variability, 10% of the stand survives the blight.

A simple chi-square calculation shows that the significance (p-value) of the separation is 4.887e-07. In lay terms: there is 1 chance in two million that this contrast happened by chance.

Thus this site, judiciously studied, and with many conditions and features identical, and dramatically different outcome, may help us to understand the nature of the sword fern blight.

Results of Soil Tests Conducted Separately on the Two SubSites

Four soil samples were collected on each subsite, to a depth of three inches. These were pooled to create two aggregate samples, one for each side, and sent to EarthFort of Corvallis Oregon for analysis.