A mirror dataset designed for hydrometeorological time series analysis in Brazil

A mirror dataset

This dataset is a mirror of official data released by Brazil's hydrometeorological public services. The dataset here is streamlined for processing and managing large volumes of data.

Data availabe

Hourly time series of automatic weather stations

Available files:

inmet_autoh_db.parquet -- database with time series of INMET automatic hourly stations;

inmet_autoh_catalog.csv -- catalog of INMET automatic hourly stations;

inmet_autoh_glossary.csv -- glossary of catalog and database of INMET automatic hourly stations;

Number of stations in catalog: 564

Start of sampling: 2000-01-01 00:00:00

End of sampling: 2025-01-01 00:00:00

Meteorological variables:

p -- Total hourly precipitation (mm);

pa_loc -- Atmospheric pressure at the station level hourly (MB);

rad -- Global radiation (kJ/m²)

temp_db -- Dry bulb hourly air temperature (°C);

temp_dp -- Dew point hourly temperature (°C);

rm -- Relative hourly air moisture (%)

wd -- Wind hourly direction (°);

wg -- Wind hourly maximal gust speed (m/s);

ws -- Wind hourly speed (m/s);

Data was retrieved from: INMET (2025). Meteorological Database for Education and Research (BDMEP) from National Institute of Meteorology (Instituto Nacional de Meteorologia - INMET). Available at: https://bdmep.inmet.gov.br/ Latest access: February of 2025.

Daily time series of conventional weather stations

Available files:

inmet_convd_db.parquet -- database with time series of INMET conventional daily stations;

inmet_convd_catalog.csv -- catalog of INMET conventional daily stations;

inmet_convd_glossary.csv -- glossary of catalog and database of INMET conventional daily stations;

Number of stations in catalog: 135

Start of sampling: 1961-01-01 00:00:00

End of sampling: 2025-01-01 00:00:00

Meteorological variables:

p -- Total daily precipitation (mm);

evp -- Daily Piché evaporation  (mm);

ist -- Total sunshine daily (h);

temp_max -- Maximum daily temperature  (°C);

temp_mean -- Compensated daily average temperature (°C);

temp_min -- Minimum daily temperature (°C);

rm_mean -- Relative air humidity daily average (%);

rm_min -- Relative air humidity daily minimum (%);

ws_mean -- Wind daily average speed (m/s);

Data was retrieved from: INMET (2025). Meteorological Database for Education and Research (BDMEP) from National Institute of Meteorology (Instituto Nacional de Meteorologia - INMET). Available at: https://bdmep.inmet.gov.br/ Latest access: February of 2025.

Quick queries

Query a station time series from the dataset with python:

# 0) Import pandas and pyarrow
import pandas as pd
import pyarrow.dataset as ds
import pyarrow.parquet as pq
# 1) Define the Parquet dataset
dataset = ds.dataset("./inmet_autoh_db.parquet", format="parquet")
# 2) Filter for a specific station id (e.g., "A001")
filtered_table = dataset.to_table(filter=ds.field("station_id") == "A001")
# 3) Convert to Pandas DataFrame and keep 'station_id' as a category
df_filtered = filtered_table.to_pandas(categories=["station_id"])
# 4) Display result
print(df_filtered.head().to_string())
# 5) Save filtered dataframe
df_filtered.to_csv("./A001.csv", sep=";", index=False)

Version log

0.0.2 -- minor problem fixed in the catalog (station_start was missing)

0.0.1 -- column name convention updated to lower case and underscores

0.0.0 -- first release