Year, month and PRODES year: date fields in every program

Abstract

Year, month, month name, PRODES year (August to July) and PRODES month in one line per program, without depending on the computer’s language. Includes Power Query and DAX for those who publish in Power BI.

Level Beginner Reading 6 min Data October 2026 Programs ArcGIS

Every monitoring dashboard needs to break dates into parts: year, month, month name and, for deforestation, the PRODES year, which runs from August 1 to July 31. “PRODES 2025” starts in August 2024.

In the old scripts, each part was a function with an UpdateCursor, repeated across several files. Here each one becomes a single line in each program.

Two month rulers. On top, the calendar year from January to December. Below, the PRODES year 2025, from August 2024 (month 1) to July 2025 (month 12), with arrows linking the corresponding months. — The PRODES year starts in August: August is month 1 and July is month 12.

The rules

The field names are in Portuguese, as in the dashboards: ano = year, mes = month, mes_nome = month name, ano_prodes = PRODES year, mes_prodes = PRODES month.

Field	Rule	2025-03-15	2025-09-20
`ano`	year of the date	2025	2025
`mes`	month of the date	3	9
`mes_nome`	month name, from a fixed list	March	September
`ano_prodes`	year + 1 if the month is August or later	2025	2026
`mes_prodes`	August = 1 … July = 12, or `(mes + 4) % 12 + 1`	8	2

% is the remainder of the division: (9 + 4) % 12 + 1 = 13 % 12 + 1 = 1 + 1 = 2.

Month name from the computer’s language

The old script used strftime("%B"), which returns the month name in the system language: “março” on a computer in Portuguese, “March” on a server in English. In a dashboard, that mixes languages. Use a fixed list of names, as in the code below.

The code

The examples use the deter layer (field view_date) of the articles’ GeoPackage (how to download).

import arcpy

fc = r"C:\dados\resultados.gdb\deter"
arcpy.management.CopyFeatures(r"C:\dados\artigos.gpkg\main.deter", fc)

meses = ("['January','February','March','April','May','June','July',"
         "'August','September','October','November','December']")

arcpy.management.CalculateFields(fc, "PYTHON3", [
    ["ano",        "!view_date!.year"],
    ["mes",        "!view_date!.month"],
    ["mes_nome",   f"{meses}[!view_date!.month - 1]"],
    ["ano_prodes", "!view_date!.year + (1 if !view_date!.month >= 8 else 0)"],
    ["mes_prodes", "(!view_date!.month + 4) % 12 + 1"],
])

CalculateFields (plural) computes several fields at once and creates the ones that do not exist. In ArcMap, use one CalculateField_management per field, with "PYTHON_9.3".

In the Field Calculator, create each field with the expression:

Field	Type	Expression
`ano`	integer	`year("view_date")`
`mes`	integer	`month("view_date")`
`mes_nome`	text	`array_get(array('January','February','March','April','May','June','July','August','September','October','November','December'), month("view_date") - 1)`
`ano_prodes`	integer	`year("view_date") + if(month("view_date") >= 8, 1, 0)`
`mes_prodes`	integer	`(month("view_date") + 4) % 12 + 1`

In the Python console, the same five fields in sequence:

import processing
G = "C:/dados/artigos.gpkg"

campos = [
    ("ano", 1, 'year("view_date")'),
    ("mes", 1, 'month("view_date")'),
    ("mes_nome", 2, "array_get(array('January','February','March','April','May','June',"
                    "'July','August','September','October','November','December'), "
                    'month("view_date") - 1)'),
    ("ano_prodes", 1, 'year("view_date") + if(month("view_date") >= 8, 1, 0)'),
    ("mes_prodes", 1, '(month("view_date") + 4) % 12 + 1'),
]
camada = f"{G}|layername=deter"
for nome, tipo, formula in campos:          # type: 1 = integer, 2 = text
    camada = processing.run("native:fieldcalculator", {
        "INPUT": camada, "FIELD_NAME": nome, "FIELD_TYPE": tipo, "FIELD_LENGTH": 12,
        "FORMULA": formula, "OUTPUT": "memory:"})["OUTPUT"]

resumo = processing.run("qgis:statisticsbycategories", {
    "INPUT": camada, "CATEGORIES_FIELD_NAME": ["ano_prodes"], "OUTPUT": "memory:"})["OUTPUT"]
for f in resumo.getFeatures():
    print(f["ano_prodes"], f["count"])

SELECT view_date,
       extract(year FROM view_date)::int                                   AS ano,
       extract(month FROM view_date)::int                                  AS mes,
       (ARRAY['January','February','March','April','May','June','July',
              'August','September','October','November','December'])
           [extract(month FROM view_date)::int]                            AS mes_nome,
       extract(year FROM view_date + interval '5 months')::int             AS ano_prodes,
       (extract(month FROM view_date)::int + 4) % 12 + 1                   AS mes_prodes
FROM artigos.deter
LIMIT 5;

-- Check: alerts by PRODES year
SELECT extract(year FROM view_date + interval '5 months')::int AS ano_prodes, count(*)
FROM artigos.deter
GROUP BY 1
ORDER BY 1;

The + interval '5 months' trick: adding 5 months to any date from August to December pushes it into the next year, while dates from January to July stay in the same year. That is exactly the PRODES year rule.

import geopandas as gpd

deter = gpd.read_file("dados/artigos.gpkg", layer="deter")
data = deter["view_date"].astype("datetime64[ns]")

MESES = ["January", "February", "March", "April", "May", "June", "July",
         "August", "September", "October", "November", "December"]

deter["ano"] = data.dt.year
deter["mes"] = data.dt.month
deter["mes_nome"] = deter["mes"].map(lambda m: MESES[m - 1])
deter["ano_prodes"] = deter["ano"] + (deter["mes"] >= 8).astype(int)
deter["mes_prodes"] = (deter["mes"] + 4) % 12 + 1

print(deter[["view_date", "ano", "mes", "mes_nome", "ano_prodes", "mes_prodes"]].head())
print(deter.groupby("ano_prodes").size())

In Power Query (Home → Transform data), use Add Column → Custom Column:

// ano_prodes
Date.Year([view_date]) + (if Date.Month([view_date]) >= 8 then 1 else 0)

// mes_prodes
Number.Mod(Date.Month([view_date]) + 4, 12) + 1

// mes_nome (fixed list, independent of the language)
{"January","February","March","April","May","June","July",
 "August","September","October","November","December"}{Date.Month([view_date]) - 1}

Or in DAX, as a calculated column of the calendar table:

PRODES year = YEAR ( dim_calendario[data] ) + IF ( MONTH ( dim_calendario[data] ) >= 8, 1, 0 )
PRODES month = MOD ( MONTH ( dim_calendario[data] ) + 4, 12 ) + 1

To sort months in PRODES-year order in charts, select the mes_nome column and use Sort by column → mes_prodes.

Checking the result

The QGIS, PostGIS and Python versions gave the same count of alerts by PRODES year:

DETER alerts by PRODES year, all classes.
PRODES year	Period	DETER alerts in the study area
2024	Jan to Jul 2024 (the download starts in 2024)	639
2025	Aug 2024 to Jul 2025	5,787
2026	Aug 2025 to Jul 2026	595
2027	Aug to Oct 2026 (ongoing)	802

The 2025 peak is 5,127 burn scars from the 2024 drought. Separating by class, clear-cut deforestation alerts (DESMATAMENTO_CR) numbered 431 in 2025 and 300 in 2026. That is why dashboards almost always separate deforestation from fire.

First and last day of the period

Two other common questions: “what are the first and last alerts in the layer?” (for the report title) and “what is the last day of the month?” (to know whether the month is closed).

	First and last date	Last day of the month
ArcGIS Pro	`min()`/`max()` over `arcpy.da.SearchCursor(fc, ["view_date"])`	`calendar.monthrange(year, month)[1]`
QGIS	`minimum("view_date")`, `maximum("view_date")`	`make_date(year(d), month(d), 1) + '1 month' - '1 day'`
PostGIS	`min(view_date)`, `max(view_date)`	`(date_trunc('month', d) + interval '1 month - 1 day')::date`
Python	`deter["view_date"].min()`	`date + pd.offsets.MonthEnd(0)`
Power BI (DAX)	`MIN()`, `MAX()`	`EOMONTH ( date, 0 )`

Updates

2026-10-29: first published, with data from INPE, MMA, IBAMA and SICAR from October 2026.

Versão em português

Reuse

CC BY 4.0

Citation

BibTeX citation:

@misc{alves2026,
  author = {Alves, Hendesson},
  publisher = {Hendesson Alves},
  title = {Year, Month and {PRODES} Year: Date Fields in Every Program},
  date = {2026-10-29},
  url = {https://hendesson.github.io/en/articles/05-date-fields.html},
  doi = {10.5281/zenodo.23780254},
  langid = {en},
  abstract = {Year, month, month name, PRODES year (August to July) and
    PRODES month in one line per program, without depending on the
    computer’s language. Includes Power Query and DAX for those who
    publish in Power BI.}
}

For attribution, please cite this work as:

Alves, Hendesson. 2026. “Year, Month and PRODES Year: Date Fields in Every Program.” In Geoprocessing Articles for Environmental Monitoring. Preprint, Hendesson Alves, October 29. https://doi.org/10.5281/zenodo.23780254.