Why JMA’s AMeDAS and Observatory Tables Have Different Fields
One portal, two table families
Japan’s official Past Weather Data Search looks like a single archive: choose a place, choose a date, then choose a table. Behind that interface, however, JMA uses separate historical table families for surface observatories and AMeDAS sites. Their page addresses commonly contain s1 and a1, respectively.
That distinction is not a minor technical detail. It helps explain why one station offers pressure and humidity while another returns only precipitation, temperature and wind, or why the available columns change when the resolution changes. A sound comparison begins by identifying the source-table family and the fields actually reported—not by assuming every dot on a national station map is equivalent.
What AMeDAS is designed to observe
AMeDAS is JMA’s Automated Meteorological Data Acquisition System, a network of automatic weather stations used for high-resolution monitoring across Japan. JMA’s current AMeDAS overview describes about 1,300 rain gauges at average intervals of 17 km. Around 840 sites also observe wind, temperature and humidity, while snow depth is observed at a smaller subset of stations in snowy districts.
Those figures reveal the important pattern: AMeDAS is not one uniform equipment package. A precipitation-only site and a multi-element site both belong to the network, but they cannot support the same analysis. Even the official AMeDAS table of display elements contains qualifications—for example, snow depth is site-dependent, and sunshine duration can be estimated rather than directly observed at certain unmanned sites.
The network’s spatial density makes it valuable for seeing local variation, but density does not create variable parity. Before comparing AMeDAS sites, a reader needs to check which elements exist at each site, for the requested period and resolution.
What the observatory tables add
JMA separately describes a surface-observation system operated at local meteorological offices and special automated weather stations. Its Surface Observation overview lists elements including surface pressure, temperature, humidity, wind, precipitation, snow, sunshine, solar radiation, visibility, cloud and weather phenomena.
Not every field is observed in the same way. Some are measured automatically; visibility, cloud and present-weather information may involve different instruments or human observation depending on station and period. This is why calling every s1 record a “manned station” can be misleading. The safer description is surface-observatory table: it identifies the portal family without making an unsupported claim about the site’s current staffing.
For historical-data users, the practical difference is that an s1 table can expose a broader meteorological context than an a1 table. Pressure and humidity, for example, may accompany temperature, precipitation and wind. But the source headings still need to be checked because station equipment and reporting practices change through time.
Why the columns change
Four factors determine what appears in a JMA historical table.
1. Network and table family
An a1 page follows the AMeDAS table layout; an s1 page follows the surface-observatory layout. The two families organize variables differently and do not promise identical columns.
2. Installed instruments
AMeDAS sites observe different combinations of elements. Snow equipment is concentrated in relevant districts, and some sites report only precipitation. Observatory sites also differ in their equipment and observation program.
3. Temporal resolution
The portal offers yearly, monthly, daily, hourly and 10-minute products, among others. An accumulation, mean, maximum and instantaneous observation answer different questions. A daily precipitation total cannot preserve the timing visible in 10-minute rows, while monthly summaries intentionally remove day-to-day detail.
4. Historical change
A long station record is not automatically homogeneous. Instruments, siting, station names, reporting intervals and table conventions can change. JMA posts corrections and notices on the search portal, and the website terms warn that content may be revised, moved or deleted. Anyone making a high-stakes or reproducible claim should retain the station identifier, dates, source URL and retrieval date.
How JMA identifies a request
The portal combines several identifiers in its table URLs:
prec_noselects a prefectural or meteorological area;block_noidentifies the station within the portal’s system; and- the page family, such as
daily_s1.phpordaily_a1.php, identifies the table structure and resolution.
These values are more dependable than a display name alone. Japanese station names can be translated or romanized in different ways, and sites with similar names may not represent the same location or operating period. A reproducible note should preserve the numeric identifiers as well as the name.
A better way to compare two stations
Suppose two stations both have daily temperature values. It is tempting to align the dates and calculate a difference immediately. A more defensible workflow is:
- Record each station’s
prec_no,block_noand table family. - Confirm coordinates, elevation and operating period.
- Check that the same variable definition and unit appear in both source tables.
- Inspect missing values and source flags rather than silently dropping them.
- Note relocations, instrumentation changes or portal notices that could affect the interval.
- Only then calculate differences or trends.
This does not make comparison impossible; it makes the limits visible. AMeDAS and observatory records are complementary precisely because they were not designed as interchangeable copies of one another.
What Climate Explorer changes—and what it does not
The JMA Explorer provides an English-language map, station table, plots and parsed rows for selected JMA records. It makes broad discovery easier and offers a convenience Excel export. The JMA station-data guide explains the exact selection and export workflow.
Climate Explorer does not replace JMA’s authoritative tables. Translation and normalization can simplify headings, but they can also hide source notation if the processed output is used alone. For evidence-sensitive work, compare the parsed values with the official page and cite JMA as the data producer. JMA’s website terms require source attribution and explain how edited content should be identified.
Frequently Asked Questions (FAQ)
What is AMeDAS?
AMeDAS is JMA’s Automated Meteorological Data Acquisition System, a nationwide network of automatic stations. Sites observe different combinations of precipitation, wind, temperature, humidity and snow.
Are all AMeDAS stations equipped the same way?
No. JMA states that some sites observe precipitation only, others add wind, temperature and humidity, and a smaller subset observes snow depth. Always check the fields for the selected station and period.
What do s1 and a1 mean in JMA table URLs?
They identify different historical source-table families: s1 for surface-observatory tables and a1 for AMeDAS tables. They help explain the page structure but do not by themselves document every instrument or staffing arrangement.
Why can the available fields change for one station?
Fields depend on table family, installed instruments, temporal resolution and historical reporting changes. A station can therefore have different coverage across variables or periods.
Where can I inspect or export the observations?
Use JMA’s Past Weather Data Search for the authoritative source tables. Climate Explorer provides an independent discovery interface and an Excel convenience export, with steps documented in the JMA station-data guide.
Data Annex
| Dimension | AMeDAS table family (a1) |
Surface-observatory table family (s1) |
|---|---|---|
| Network context | Automatic regional observation sites | Surface observatories and related surface-observation sites |
| Common elements | Precipitation; temperature, wind and humidity at equipped sites; snow at a subset | Can include pressure, temperature, humidity, wind, precipitation, snow, sunshine and other surface-observation elements |
| Main comparison risk | Assuming every site has the same instrument set | Assuming every field or observation method is constant across sites and time |
| Required check | Fields, dates, units and source flags for the selected site | Fields, dates, units, observation method and source flags for the selected site |
| Portal component | Interpretation | Reproducibility note |
|---|---|---|
prec_no |
Prefectural or meteorological area code | Retain it with the station name. |
block_no |
Station identifier used by the portal | Use it to distinguish similarly named sites. |
a1 / s1 |
Historical table family | Do not translate it into a claim about current staffing. |
| Resolution | Yearly, monthly, daily, hourly, 10-minute or another offered table | Preserve the aggregation and source unit. |
| Date selection | Period requested from the portal | Record the time basis and retrieval date. |
Data Sources
- JMA Past Weather Data Search: official station, date and table-selection portal.
- JMA AMeDAS overview: official description of the automated network, observation elements and current network summary.
- JMA Surface Observation overview: official description of surface stations and observed elements.
- JMA Weather Observation overview: official context for automatic and visual surface observations.
- JMA website terms of use: attribution, reuse and revision conditions.