One hundred days into the nuclear emergency here, Japan is learning that danger doesn’t recede in tidy concentric circles centered on the damaged nuclear plant.

As officials bolster efforts to map the nuclear fallout with daily and sometimes hourly readings, they have found that radioactive particles concentrate in random hot spots — curling with the wind, collecting along mountainsides and raising fresh problems for residents who thought they were a safe distance from danger.

Japan’s evacuation map now resembles not a circle but a paw print, with a growing number of finger-like projections. On Thursday, Chief Cabinet Secretary Yukio Edano recommended the evacuation of several emerging hot spots beyond the government-ordered 12.5-mile evacuation zone around the Fukushima Daiichi plant. These hot spots — some as far as 35 miles from the facility — could receive more than 20,000 microsieverts of radiation in a year, surpassing the internationally recognized limit for adult exposure, Edano said.

Instead of basing its evacuation recommendations on simple geography, the government is now opting for a hyper-detailed evaluation that underscores the randomness of radioactive patterns. On Saturday, the government said new hot spot areas would be analyzed on a house-by-house basis, a tacit acknowledgment that the formula for projecting long-term exposure relies too much on estimation and should not be the sole factor in determining whether residents stay or leave.

As Ohara residents like Ichiro Tani, 62, and Toyoaki Matsushita, 54 — longtime friends — now know, estimating one’s radiation risk is a fraught undertaking. In their mountainside town, a cluster of 120 households about 19 miles from the plant, levels fluctuate drastically within a few blocks.

One resident, Toshihiko Nakano, ordered a Ukraine-made dosimeter and walks every Saturday from house to house, recording the measurements.

In one front yard: 3.65 microsieverts per hour.

In another: 11.3 microsieverts per hour.

The local government began paying serious attention to this area only on May 1. That’s when Minamisoma, the closest city to Ohara, began daily radiation monitoring, with two workers making eight-hour driving tours, stopping at 36 predetermined places.

Every afternoon about 1:30, the local government crew arrives at an abandoned gravel lot just a stone’s throw from Tani’s house. On Thursday, as Tani and Matshushita smoked cigarettes nearby, the workers held a dosimeter one centimeter above the ground, then recorded the digits. They repeated the process, holding the dosimeter one meter above the ground.

“What have we got today?” Tani asked.

“It’s even higher today than normal,” one of the workers said.

The number, taken from waist-level: 3.76 microsieverts per hour.

“Oh, pretty big,” Tani said.

For Tani and his neighbors, these numbers no longer seem like industry-speak abstractions. But they have struggled to put the figures into context — a common problem during these last months as the government has issued reams of data but little practical information.

Take any of these hourly numbers, Tani knows, and you can easily scare yourself. Multiply by 24 (for hours in a day), then multiply by 365 (for days in a year), and the number always soars above the 20,000 microsieverts per year.

“But that’s the simple way to calculate it,” Tani said. “I don’t know the formula.”

Stuck with an estimate

Authorities at Japan’s education and science ministry devised a formula to turn hourly rates into yearly ones. Their formula, spokesman Hirotaka Oku said, assumes that a person spends eight hours daily outdoors and the remaining 16 hours in a “wooden structure,” and that indoor radiation levels decrease by 60 percent.

But here’s where things get more complicated. During the first few weeks of the crisis, there was scant information about what, precisely, was happening. With power out and machinery malfunctioning at Fukushima, Japan’s nationwide radiation monitoring system did not work as designed, a recent government report said. Only last month did Japan confirm the full meltdown of three reactors at the plant. It also recently doubled its estimate of the radioactive material released during the accident.

Japanese authorities feel comfortable applying their formula to dates since May 25 — radiation levels have held steady in recent weeks — but they are forced to estimate radiation exposure during the first 74 days of the crisis, which started with an explosion on March 12. The government projects that anybody in Tani’s neighborhood received 6,600 microsieverts during that time.

Its projection for total annual exposure in Ohara: 23,800 microsieverts.

That’s far less than the 250,000 microsieverts allowed for emergency situation nuclear workers in Japan. But it is the equivalent of almost 500 chest X-rays or 120 New York-to-Tokyo flights.

Pondering the numbers

These days, Tani hears conflicting advice. His wife talks about wanting to leave. His closest neighbors evacuated one day after the disaster and haven’t returned.

Sometimes, Tani plays with the numbers in his head. He left town for two weeks in March, so maybe he escaped the worst fallout. Indoor exposure rates can vary depending on the type of roof one’s house has. Plus, who stays outdoors eight hours a day?

Edano said Thursday that Japan would try to be flexible for those in hot spot regions. The country wouldn’t force evacuations, just recommend them, he said.

“But it is natural for residents to feel anxious,” Tani said. “If you have the right mind, you don’t stay here.”

Special correspondents Sachiko Iwata and Akiko Yamamoto contributed to this report.