Fix Log

Corrected the Koshu Kaido Manhattanhenge alignment dates from "May 4 and August 7" to "May 5 and August 8." The original dates were based on a preliminary calculation that did not account for the exact street azimuth. David Park recalculated using the precise alignment of 255.3 degrees and the NOAA solar position algorithm. The new dates are accurate to plus or minus one day depending on atmospheric refraction. Reported by a reader who stood on the street on May 4 and found the sun slightly off-center.

Updated the cumulonimbus frequency from "15 to 20 summer days" to "20 to 25 summer days" based on JMA lightning strike data for the Tokyo region from 2020 to 2024. Yuki Mori cross-referenced our field observations against the official records and found we had undercounted days with distant cumulonimbus visible from our observation points but not producing local lightning. The text now reflects both local and distant thunderstorm days.

Updated the Shinjuku Southern Terrace SVF from 0.15 to 0.12 after David Park conducted a detailed angular measurement using a digital clinometer from multiple points on the plaza. The original estimate was based on satellite imagery analysis, which overestimated the visible arc due to building setbacks that are not visible from ground level. The text and data have been updated.

Added documentation of the Brocken spectre observed on November 12, 2024, from Mt. Takao. Hana Suzuki's photograph was verified against EXIF data and weather conditions. The fog layer at 300 meters, clear summit conditions, and solar elevation of 8 degrees all match the required geometry for a Brocken spectre. This is our third documented spectre from Takao and the most vivid to date.

Corrected the Mt. Takao Bortle classification from "Bortle 4 to 5" to "Bortle 5." David Park's remeasurement with a newly calibrated SQM-LU gave a consistent reading of 20.7 MPSAS, which falls squarely in the Bortle 5 range. The original classification was optimistic due to meter drift that was corrected in the 2024 calibration. The sky is still the darkest accessible by train from central Tokyo.

Updated the Shibuya Sky operating hours and pricing. The deck now closes at 10 PM (previously 9:30 PM) and the adult admission has increased from 1,800 yen to 2,000 yen. We also added a note about the weekday morning recommendation based on our own visits finding significantly smaller crowds before 10 AM.

Corrected the altitude range for cirrus clouds from "7,000 to 13,000 meters" to "6,000 to 12,000 meters." Yuki Mori identified that our original range extended too high for Tokyo's latitude, where the tropopause is typically at 11,000 to 12,000 meters. Cirrus at 13,000 meters would be in the stratosphere at 35 degrees north, which is physically possible but extremely rare. The corrected range reflects typical Tokyo cirrus altitudes.

Clarified that our cloud cover data comes from model output, not direct satellite pixel measurement over each location. A reader pointed out that the ECMWF model grid cell for Odaiba also includes a portion of the bay, which means the reported cloud cover may not match the visual experience from the beach. We added explanatory text to the Sky View page and this method page to make this limitation clear.

Added documentation of the sun corona observed on May 22, 2024. This was our first clear sun corona sighting in three years of observation. The phenomenon was visible at 3 PM from Yoyogi Park through a thin altocumulus layer. Photographic evidence with neutral density filtering confirmed two visible rings of color. The event lasted approximately 8 minutes before the cloud thickened and the corona dissolved.

Updated the Japan field station address from "2-3-8 Aomi" to "2-4-10 Aomi." We moved to a larger space within the same neighborhood to accommodate our growing equipment inventory. The phone number remains the same. The old address now redirects to the new one.

Corrected the Yoyogi Park area from "121 acres" to "134 acres." The original figure came from an outdated Tokyo Metropolitan Government brochure. The current official figure, confirmed with the Parks and Gardens Association, is 54.1 hectares, which converts to approximately 134 acres.

Removed the reference to "seeing the Andromeda Galaxy with the naked eye from Mt. Takao." While the galaxy is visible in binoculars from Takao, the Bortle 5 sky brightness makes naked-eye detection marginal at best and requires exceptional eyesight and dark adaptation. We overstated the visibility in our enthusiasm. The text now accurately describes it as "a fuzzy oval in binoculars."

Corrected the Tokyo heat island temperature differential from "3 to 5 degrees" to "2 to 4 degrees Celsius." Yuki Mori's literature review found that recent studies using more comprehensive sensor networks have revised the estimate downward. The 2-to-4-degree range reflects the current scientific consensus for the average annual differential. Peak summer differentials can still reach 5 degrees or more on specific days.

Added the noctilucent cloud sighting from Mt. Takao at 3:45 AM. Hana Suzuki was the primary observer. Five other observers reported the same phenomenon from separate locations. We have no photograph — the light was too faint for her camera settings. We classify this as an unphotographed but multiply-witnessed observation. The mesospheric altitude, the blue-silver color, and the wavy pattern are all consistent with noctilucent clouds at Tokyo's latitude during the NLC season.

Corrected David Park's bicycle survey distance from "350 kilometers" to "400 kilometers." The original figure was an estimate. David later compiled his actual route logs and found the total was closer to 400 kilometers due to repeat measurements in several wards and a planned extension into the western suburbs that was not originally included in the estimate.

Corrected the golden hour duration in summer from "about 80 minutes" to "72 minutes." The original figure was rounded up from a rough calculation. David Park recalculated using the exact solar elevation angles for Tokyo's latitude on the solstice dates, giving 71.8 minutes. We round to 72 for readability. The winter figure of 48 minutes remains accurate.