How Does Rainfall Differ Between Bali’s Coastal and Inland Areas?

The most striking figure in BMKG’s February 2026 observations is 1,052.0 mm recorded in Gerokgak, western Buleleng, across 21 rainy days. Yet Gerokgak did not have the highest number of days with rain. That distinction went to Bangli subdistrict, where rain was recorded on all 28 days of February. Both figures appear in the March 2026 bulletin issued by BBMKG Region III. They describe a particular month, not the usual climate of northern Bali or a ranking of places where daily life is hardest.

A rainfall total measures accumulated precipitation. One millimetre is the equivalent of one litre falling on a square metre of level surface; it says nothing about how that water subsequently drains from a property. A monthly reading of 300 mm could be associated with several intense events or rainfall distributed across many days. Neither explanation can be established from the total alone. BMKG therefore publishes rainy-day information separately. The Indonesian label hari hujan refers to days on which rain was recorded, but the bulletin’s count should not be turned into hours of rain or assumed to use an unstated threshold. Someone who leaves home at 7 a.m. needs information that a monthly total cannot supply.

March offers an especially useful counterexample within Denpasar. The April 2026 BBMKG bulletin places both Denpasar Timur and Denpasar Barat in the 151–200 mm rainfall band for the observed month of March. Its separate rainy-day analysis places Timur above 20 days and Barat in the 11–20-day band. In other words, the same range of accumulated rainfall does not imply the same range of days on which rain was recorded. Both are regional map categories, not measurements taken at two particular houses. The distinction matters when comparing areas for commuting, and it also shows why an address-specific forecast cannot be inferred from these maps.

A third number often appears beside rainfall totals: sifat hujan, the percentage of the locally expected amount. Under BMKG’s published terminology, above normal means more than 115% of the relevant normal, normal covers 85–115%, and below normal is less than 85%. An above-normal month in one subdistrict may still produce fewer millimetres than an ordinary month elsewhere. The reference is a local climatological normal, not a competition between districts; its actual baseline period must be checked in the product being quoted. Nor should an extreme daily reading be confused with a monthly total. In February 2026 BBMKG reported the highest single-day rainfall, 216.9 mm, in Kuta, while the highest monthly station accumulation was in Gerokgak. Those are different measurements and support different conclusions.

“Ubud versus the coast” sounds like a straightforward weather comparison, but it is not how Bali’s official observations are organised. BMKG works with stations, rainfall posts, mapped subdistrict areas and season zones. A tourism label can cover very different terrain, while an administrative subdistrict may fall across several rainfall bands. The February and March 2026 maps from BBMKG Region III make those distinctions visible. Neither document establishes a rainfall figure for a weather station in central Ubud, so figures from nearby Payangan or Tampaksiring must be described as inland Gianyar context, not as Ubud’s measured total.

Consider the inland–urban contrast first. In February, Payangan was mapped at 401–500 mm and Tampaksiring above 500 mm. Both Denpasar Timur and Denpasar Barat were also above 500 mm, as were Kuta and Kuta Selatan in Badung. In March, Payangan moved into the 301–400 mm band and Tampaksiring into 151–200 mm. Denpasar’s two mapped sectors were in that same 151–200 mm class; Kuta was also at 151–200 mm, while Kuta Selatan was at 201–300 mm. The relative positions of Payangan and Tampaksiring changed between the two observed months. A categorical statement that inland Gianyar is wetter than the southern urban fringe throughout the year would go beyond these observations.

Northern Bali is equally varied. February’s Gerokgak mapping covered parts of the subdistrict in the 151–200, 201–300 and 401–500 mm categories. A separate rainfall post in Gerokgak recorded the exceptional 1,052.0 mm monthly total; that station figure must not be painted across the entire subdistrict. In Tejakula, portions were shown at 301–400, 401–500 and above 500 mm. March again divided Gerokgak among several bands, from 101–150 up to 401–500 mm, while Tejakula included areas in the 201–300 and 401–500 mm classes. Even within Buleleng, Busung Biu was above 500 mm in February and in the 401–500 mm band in March. Its inland setting also illustrates why a regency name is a poor substitute for “north-coast climate”.

The uplands present a similar problem at a finer scale. In Bangli regency, parts of Kintamani fell into February categories of 401–500 mm and more than 500 mm. March brought mapped portions at 301–400, 401–500 and above 500 mm. Baturiti in Tabanan had parts in three February classes—201–300, 301–400 and above 500 mm—and two March classes, 101–150 and 201–300 mm. Selecting one of those values as the rainfall of “Kintamani” or “Baturiti” would conceal the variation the official map is showing. Bali’s climate office notes that topography is a local influence on rainfall, but these reports cannot support a precise altitude-based formula for individual properties.

The table below therefore uses seven subdistricts for which each bulletin gives a clear single mapped rainfall band in both months. Gerokgak, Tejakula, Kintamani and Baturiti remain important to the interpretation precisely because one number would misrepresent them. Read each entry as a dated map classification, not as a direct station total or a block-by-block estimate. February and March are also adjacent observed months, so the table illustrates spatial differences and changes within one part of the seasonal cycle; it does not establish a full wet-season versus dry-season ranking.

Observed monthly rainfall bands in the BBMKG III February and March 2026 maps, in mm/month. March replaces July because district-level July observations were not substantiated. These adjacent months are not a wet-versus-dry-season benchmark. “>500” has no stated upper bound. Payangan is not a Ubud station.

Someone moving for a school term is asking a different question from someone renting for six weeks in July. Both need calendar-specific evidence. A long-term rainfall normal describes what has been typical over a defined reference period; a monthly bulletin reports what actually fell in one month; a seasonal outlook offers an estimate of what may happen next. These categories should not be blended. For example, the Gerokgak station total of 1,052.0 mm belongs to February 2026, while the highest March 2026 monthly accumulation reported by BBMKG was 595.3 mm at Sukasada, Buleleng, with 19 rainy days. Neither station maximum is a sensible “expected rainfall” figure for a forthcoming rental stay.

The February–March table is deliberately limited. It establishes genuine changes in mapped monthly totals across several places, but those are consecutive months within the broader seasonal transition. It would be misleading to label the March column “dry season” or to treat the pair as the full annual pattern. A July or August stay calls for records from July and August, ideally across more than one year and from geographically relevant observations. An older forecast for July is not an observed July rainfall map. Equally, a single rainy July in a historical record cannot decide whether July will be dry at a specific villa next year. Comparing like months across years is a different exercise from comparing different subdistricts during the same month.

Bali is covered by 20 BMKG seasonal zones, or Zona Musim (ZOM). Their boundaries reflect seasonal rainfall patterns and need not follow regency or subdistrict lines. BMKG also divides a month into three planning periods called dasarian: days 1–10, 11–20 and 21 to month-end. A statement about the expected onset during one dasarian is therefore an outlook for a seasonal zone, not a daily forecast. This matters on an island where neighbouring homes may occupy different terrain and where one regency can cross several seasonal classifications. The national seasonal outlook is valuable for the wider climate picture, but nationwide percentages of zones must never be presented as Bali-specific probabilities.

Stasiun Klimatologi Bali announced its 2026/27 rainy-season onset bulletin on 13 October 2026. The published notice establishes that a local outlook exists; it does not, by itself, substantiate dates for individual zones, so this page does not assign them. Where a forecast uses a climatological normal, its reference years must be taken from that product rather than borrowed from a glossary or an older seasonal report. For a December-to-February household, the next useful evidence is a set of matching-month totals and rainy-day counts, followed by the latest zone-specific seasonal guidance. For a remote worker choosing July and August, a longer series from those months has greater value than a February record. The calendar narrows which evidence to read; it does not remove the need to inspect the address.

A mapped rainfall band is not a property inspection. It may describe part of the subdistrict containing a house, but it does not reveal the slope of its driveway, the condition of the roof, or the hours when rain usually reaches that street. The first useful distinction is geographical: where is the building, which kecamatan contains it, and how close are the relevant observation sites? The answer may differ from the name on a holiday-rental listing. A home advertised as “Ubud” cannot automatically inherit numbers mapped for Payangan or Tampaksiring. Those inland Gianyar areas provide context, not an invented central-Ubud weather-station reading.

Now consider a parent who has to reach school by a fixed time every weekday. The March 2026 data offer a warning against relying on one number: Denpasar Timur and Denpasar Barat had the same mapped 151–200 mm rainfall range, but different rainy-day categories. That difference may prompt a closer look at travel routines, yet it still does not tell us whether rain fell during the school run. A rainy day need not be a day of disrupted traffic, and a high monthly total cannot establish that a road floods. The evidence that answers the parent’s question is practical and local: the actual morning and afternoon journey, a workable alternative if a scooter is unsuitable, and dated observations of the route after heavy rain. No monthly bulletin can substitute for that.

A person working from home faces a different set of unknowns. Rainfall totals do not demonstrate that a bedroom will feel damp, that an office will remain comfortable, or that water enters through a roof. These are building-condition questions. Photographs and short videos taken after a downpour can show the entrance, gutters, roof junctions, ceilings and workspace. A landlord or manager can also explain how the property is ventilated and who is responsible for investigating leaks. Such evidence has to identify the actual home and the date. A pristine photograph from one afternoon is a useful observation of that afternoon, not a guarantee for the whole tenancy.

The best use of regional rainfall evidence is to decide where a more detailed property check is worth the effort. If a mapped area registered high rainfall during the months relevant to the tenancy, that is a reason to ask more pointed questions about maintenance and access. It is not proof that the house has a defect. Equally, a relatively modest monthly rainfall band says nothing conclusive about how a particular driveway handles a short, intense shower. An official climate bulletin, the manager’s account, and direct observations each answer different questions. Their reliability improves when their dates, locations and limitations remain visible; combining them into a made-up “villa weather score” would only disguise uncertainty.

This is why two superficially similar villas can call for different decisions. One may be judged mainly by its fixed-time commute and transport options; another by the quality of the indoor workspace and the building’s response to heavy rain. Neither scenario permits deductions about mould, flood frequency, impassable roads or repair costs from rainfall totals alone. Flood exposure, hillside stability and site drainage need their own local evidence, beyond the subject of this rainfall comparison. The sensible final step is therefore not to demand a precise street forecast from BMKG, but to match the regional, dated climate picture with dated evidence of the particular address and the week someone intends to live there.