Japan Dive Site Visibility Ranking — 46,000+ Observations Compared

Water visibility is one of the most important factors in a great diving experience. Yet visibility reports from individual dive shops vary in measurement methods and frequency, making cross-site comparison nearly impossible.

This article analyzes over 30,000 actual visibility observations collected by our site from 2006 to 2026, covering more than 30 dive sites across Japan. Beyond simple averages, we present median, standard deviation, 95% confidence intervals, and monthly observation counts to clarify both the reliability and limitations of the data.

Data Collection & Methodology

Data Sources

Data is automatically collected via web scraping from dive shop blogs and log pages across Japan. Target platforms include WordPress, ExBlog, Livedoor Blog, Ameblo, Wix, and custom CMS sites. Visibility values (e.g., "visibility 10-15m") and water temperature are extracted from each post using regular expressions and stored in a structured SQLite database.

Visibility Calculation

Visibility for each observation is calculated as the average of the reported minimum (vis_min) and maximum (vis_max). For example, "visibility 10-15m" yields 12.5m. Single values (e.g., "visibility 10m") are used as-is. The site average is the arithmetic mean of all observation-day values.

Ranking Criteria

• Only sites with 100+ observations are included (32 sites qualify)
• Observation periods vary by site (longest: Echizen from 2006; shortest: Koganezaki from 2024)
• Duplicate records (same site, same date) are excluded via UNIQUE constraint

Data Limitations & Known Biases

Visibility Map — Nationwide Overview

Explore each dive site's location and visibility on the map below. Circle size and color represent visibility. Switch months to compare seasonal distribution changes.

Visibility Ranking (with Statistical Metrics)

The table below shows visibility statistics for each site. In addition to the mean, we include the median (robust to outliers), standard deviation (spread), and 95% confidence interval (estimated range of the true mean). Narrower confidence intervals indicate more reliable data.

How to Read the Statistics

• Mean vs. Median: When the median is lower than the mean, high-visibility outliers may be pulling the average up. A small gap between the two indicates a symmetric distribution and a more reliable representative value.
• Standard Deviation (SD): A larger SD means greater variability in visibility. Compare the SD values across sites in the table above.
• 95% Confidence Interval (CI): More observations lead to narrower intervals. Sites with CI width under 1m have high precision in their mean estimates.

Monthly Observation Count — Visualizing Seasonal Bias

To assess ranking reliability, we need to verify that data is evenly distributed throughout the year. The heatmap below shows monthly observation counts for each site. Lighter months have fewer data points, meaning visibility estimates for those months are less reliable.

For example, Echizen has only 23 observations for November through February in our database (out of 2,652 total), so its annual average of 8.9m effectively reflects March-October data only. This is likely due to some facilities closing in winter (ECHIZEN LOG) and rough Sea of Japan winter conditions (conjecture). In contrast, Izu Oceanic Park (3,151 total) and Kushimoto (3,168 total) have 200+ observations per month, making their annual averages highly reliable.

Monthly Ranking — Apples-to-Apples Comparison

Annual averages are affected by observation periods and seasonal bias. Comparing within the same month provides a fairer ranking. For example, in our database, Izu Oceanic Park in January averages 18.6m (278 observations) — its highest visibility month. The high winter visibility on the Izu Peninsula is associated with the approach of the Kuroshio Current (JMA: Kuroshio, Diving School Beyond). In August, Ishigaki averages 21.0m (212 obs.) and Kerama 20.9m (132 obs.), while Izu Oceanic Park drops to 12.3m (279 obs.).

Detailed Analysis of Top Sites

Yakushima (avg. 24.8m) — Data Limitations Apply

Ranked #1 by average in our database, but with only 156 observations — an order of magnitude fewer than other top sites (Ishigaki 1,473; Kerama 1,533). The 95% confidence interval in the table above is wide, indicating limited statistical precision.

The monthly observation heatmap above shows data concentrated in specific months. The small sample size and seasonal bias may inflate the average above the true value (conjecture).Yakushima is directly influenced by the Kuroshio Current, and visibility over 30m has been reported under good conditions (Yakushima Diving Points). However, statistical reliability of this ranking is inferior to data-rich sites like Ishigaki and Kerama.

Ishigaki (avg. 20.5m) — Most Reliable Data

With 1,473 observations in our database, Ishigaki shows a narrow confidence interval as seen in the table above. Monthly averages range from 19.1m (November) to 21.9m (March) — remarkably stable year-round. It consistently ranks near the top in the monthly ranking table above.

Kerama (avg. 19.4m)

With 1,533 observations in our database, averaging 19.4m. Monthly averages peak at 21.4m in July (139 obs.) and bottom out at 18.1m in January and February (116 and 119 obs. respectively), demonstrating high stability year-round.

Regional Monthly Rankings

The national ranking can obscure differences within regions. Below, we present monthly rankings by area, making it easier to identify the best dive site for each season within a given region.

Conclusion & Future Work

Our database analysis reveals clear regional and seasonal differences in dive site visibility across Japan, demonstrating that simple annual averages are insufficient for fair comparison. Limiting to sites with 1,000+ observations, Ishigaki (20.5m, 1,473 obs.) and Kerama (19.4m, 1,533 obs.) show the highest average visibility and stability. Izu Oceanic Park averages 18.6m in January (278 obs.), approaching Okinawa-level values during winter.

Future work includes collecting measurement depth data, quantitatively correcting for fair-weather bias, and standardizing comparisons across equivalent time periods. We continue to expand data collection and refine analytical methods.

Note that visibility is only one aspect of a great dive. For example, Osezaki Bay (7.6m average in our database) is renowned as a "macro paradise" (Marine Diving web), and Ito (16.0m) is famous for its "Shark Scramble" — massive schools of banded houndsharks (Ito Diving Service BOMMIE). Please use visibility data as one of many factors in planning your dives.