The Wardrobe No One Measured

Open a wardrobe you have kept shut for weeks, and before you see anything, a cold, heavy smell greets you, one you know well. That smell is not a flaw in the wood or the fabric. It is an instant reading of a number no one measured: relative humidity. That number decides the fate of walls, roofs and furniture, and the comfort of the people who live behind them. Here lies the paradox. Chachapoyas, a city in northern Peru at roughly 2,490 meters above sea level and surrounded by cloud forests, has, according to our review of the peer-reviewed literature, no independent, published relative humidity record of its own. Where, then, do the numbers come from on which everything built there rests? The answer exposes a side of architecture that lectures do not teach.

The Smell That Precedes the Number: A City Without a Humidity Record

When we searched the indexed studies for climate data on Chachapoyas, we found one meteorological station known to the literature, but no series of measured relative humidity from the city itself. What exists are gridded climate layers for the region and short field campaigns in other Peruvian cloud forests. The researcher Altea, who studied how farmers and institutions perceive climate change in the Amazonas region, describes the problem plainly. Near the communities she studied, she found only two weather stations, installed by the local university in 2012, and she noted that the national meteorological service’s historical records for the region are, for the most part, unavailable.

The wider picture is not much different. In their review of climate data sources in the Andes, Condom and colleagues counted about 2,236 stations in the Peruvian network, only 315 of them between 2,001 and 3,000 meters, the elevation band in which Chachapoyas sits. The World Meteorological Organization has received reports from only 55 of them. A figure like that means that anyone designing a building at these altitudes works, for the most part, without a scale.

The consequences for architectural practice are direct. The basic decisions of envelope design, such as wall thickness, the placement of thermal insulation, how spaces are ventilated and how rainwater is drained, presuppose knowledge of the humidity a building will endure over decades. Where measurement is absent, estimation, imitation or solutions copied from other regions take its place. Still, the one known station offers at least something, and what it says deserves a hearing.

A Station at 2,490 Meters Tells Half the Story

The research of Lavado Casimiro and colleagues shows that the Chachapoyas station (at 6.203 degrees south and 77.881 degrees west, at an elevation of 2,490 meters) belonged to a network of 77 stations across the Huallaga and Ucayali basins, used to analyze temperature and rainfall trends between 1965 and 2007. The station supplies monthly temperature series, mean, maximum and minimum, but no humidity series. The researchers concluded that the region has warmed by roughly 0.09 degrees Celsius per decade, with stronger trends in summer maximum temperatures in the Andean zones and no significant trend in annual rainfall over the same period. They also documented the influence of El Nino and of tropical Atlantic temperatures on fluctuations in temperature and rainfall.

In another study by the same team, on rainfall and runoff across Peru’s basins, they found that rainfall typically falls below 1,500 millimeters a year above 2,000 meters, and that the inter-Andean sector around Chachapoyas receives about 1,000 millimeters a year. The study also recorded a decline in runoff after the mid-1980s.

An increase of nine-hundredths of a degree Celsius per decade may look like a number without weight. But a building erected today may stand for a century, and that number accumulates across its lifetime. Reading temperature is tied to reading humidity because warmer air carries more vapor, which changes the conditions for condensation on the cold surfaces inside walls. This is an engineering inference from the physical relationship between temperature and humidity, not something the study demonstrated. The upshot is that the absence of a significant trend in annual rainfall does not mean the climate a building faces is constant, because a single year contains two radically different seasons. Consider how long the wet season lasts.

A Seven-Month Rainy Season Writes the Fate of the Walls

The study by De la Cruz and colleagues, built on daily records from 377 stations between 1981 and 2019, offers a precise picture of the rhythm of rain in Peru’s northeastern highlands, the region in which Chachapoyas lies. The rainy season begins on average on October 15 and ends on May 6, about 204 days, and accumulates roughly 738 millimeters. More important, rainfall accumulated over the season rose by a significant 44.9 percent between 1981 and 2019.

This picture is consistent with what Altea reported for the nearby Utcubamba valley, where the wet season runs from October to May and is followed by a semi-dry season from June to September, with a risk of drought in July and August. Lake Pomacochas in Bongara province, studied by Bush and colleagues, offers measured readings of the modern climate: a mean temperature of 15.1 degrees Celsius, annual rainfall of 881 millimeters (average for 1969 to 1973) and a four-month dry season from May to August. The numbers differ from one site to another and from one measurement method to another, and that is itself a lesson: there is no single “Chachapoyas number,” only a range within which the truth moves.

For the architect, this means that a wall in this region lives a double cycle: long months of continuous wetting, then a shorter window for drying. A good building is one that lets its moisture out in the dry season as fully as it took it in during the wet season. If the envelope materials trap water or slow its exit, the dry months will not suffice to restore balance. This is a design inference from the documented rhythm of the climate, not a verdict issued by the studies. But rain alone does not tell the whole story, for here the clouds themselves move in.

When the Cloud Becomes a Permanent Guest on the Facade

The study by Sales and colleagues of Lake Progreso in the Amazonas region, at 2,013 meters, describes a modern climate with a mean temperature of 17.2 degrees Celsius, rainfall of about 1,126 millimeters and a rainy season from October to April. It discusses the high humidity, fog and cloud immersion that characterize the cloud forest belt between 1,500 and 2,500 meters.

Studies at other sites in Peru clarify what these words mean. In the Yanachaga-Chemillen forest of central Peru, at elevations of 2,468 and 2,815 meters during 2003 and 2004, Gomez-Peralta and colleagues measured rainfall, fog water and the water retained on leaves. They found rainfall ranging from 2,222 to 2,753 millimeters a year, fog water that exceeded rainfall in some weeks, and canopy interception that took between 7.7 and 29.6 percent of the water. Rapp and Silman measured temperature and relative humidity every ten minutes along a slope between 1,500 and 3,500 meters in the Kosnipata valley near Manu. Humidity stood above 95 percent in most measurements during the rainy season, and the researchers also calculated water vapor and vapor pressure deficit, which is the air’s capacity to dry things.

These sites are not Chachapoyas, a limitation that must be stated. But they serve as a methodological model for understanding life inside a cloud. The architect may ask: if tree leaves catch fog water and retain part of it, what happens to a building’s facade and roof? And if the air hovers near saturation for weeks on end, the vapor pressure deficit, which measures the air’s ability to dry surfaces, falls so low that natural drying becomes very slow. This is a design reading of what the studies reveal about the physical behavior of cloud air, not a claim about the measurement of any particular building. And because direct measurement is missing, digital substitutes had to try to fill the gap.

Numbers from the Sky: When Satellites Replace Stations

The closest the peer-reviewed literature comes to humidity in the Amazonas region is the study by Salas Lopez and colleagues on land suitability for coffee cultivation. The researchers did not obtain humidity from ground stations. They calculated it from FLDAS reanalysis data on the Google Earth Engine platform. For the remaining variables they relied on WorldClim 2.1 layers for 1970 to 2000 at a resolution of about one kilometer. They set an optimal humidity range for coffee between 70 and 90 percent, and recorded that mean annual temperature in the Andean sector of the region ranges from 7.4 to 19.8 degrees Celsius. They provided statistics at the provincial level, including Chachapoyas, Luya, Bongara and Utcubamba.

Rainfall, for its part, is covered by the PISCO gridded product described by Aybar and colleagues, at a resolution of 0.1 degrees (about ten kilometers) since 1981. It rests on a national network of 945 rain gauges, of which only 441 series proved usable after quality review and gap filling. Nearly half the gauges fell by the wayside.

This poses a dilemma for anyone working in a narrow mountain valley. A ten-kilometer cell may contain an exposed slope, a sheltered valley floor, a channel of cold air and a forest edge, each living a different humidity. The cell’s average value may describe no point within it with precision. This is an observation about the limits of gridded products in general, not an accusation against these studies, which acknowledge their own limits. But the digital numbers, constrained as they are, are what is available today. The larger question is what the climate will look like when buildings are tested against it decades from now.

Tomorrow’s Climate Will Be Tested on Buildings Designed with Yesterday’s Numbers

The study by De la Cruz and colleagues offers future climate projections (CMIP6 under the SSP2-4.5 and SSP5-8.5 pathways) for the onset and end of the rainy season and for its rainfall amount in Peru’s northeastern highlands. As noted above, accumulated rainfall in this region rose by 44.9 percent over the past four decades. If that trend continues, a building designed today on the old average figure will be fragile in the face of tomorrow.

On a wider scale, Potter and colleagues combined a high-resolution simulation with thirty downscaled climate models for the Peruvian Andes. Under the RCP8.5 scenario by the end of the century, they projected warming between 3.6 and 4.1 degrees Celsius, an increase in rainfall of about 12 percent, a rise of roughly 75 percent in precipitation on very wet days, and intensifying atmospheric drought. Their study focuses on the Cordillera Blanca and Vilcanota-Urubamba, not directly on Amazonas, but it is the most recent reference for the climate change framework of the northern and central Andes, and it carries implications for the region.

This means a building in Chachapoyas may face both heavier rain on its stormy days and sharper dry spells. An envelope that withstands the average is not enough; what is needed is one that withstands the extremes in both directions. Here the cost of the missing humidity record multiplies, because future models need a present-day baseline against which to be measured. Without a baseline, the uncertainty passes from the present into the entire future.

The Wardrobe No One Measured

Let us return to the closed wardrobe and the smell that arrives before sight. You do not need a hygrometer to know that something has gone wrong; your body reads the air before any instrument does. Buildings have no such sense, and they spend their long years behind walls receiving whatever vapor the air carries, with no one telling them how much. In a city that stands 2,490 meters high and is ringed by clouds, architects and builders design, and people live, in houses whose exposure to humidity has never been precisely known. We have one known station that measures temperature, digital grids with ten-kilometer cells, and studies from neighboring forests from which we borrow understanding. So who bears the responsibility for building against a humidity no one has measured? And is it enough to learn from the smell of the wardrobe after things have rotted, or has the time come for measurement to become part of the design of every building from the very first moment?

References

Lavado Casimiro, Waldo S., et al. “Trends in Rainfall and Temperature in the Peruvian Amazon-Andes Basin over the Last 40 Years (1965-2007).” Hydrological Processes, 2012.

Lavado Casimiro, Waldo S., et al. “Basin-Scale Analysis of Rainfall and Runoff in Peru (1969-2004): Pacific, Titicaca and Amazonas Drainages.” Hydrological Sciences Journal, 2012.

Salas Lopez, Rolando, et al. “Land Suitability for Coffee (Coffea arabica) Growing in Amazonas, Peru: Integrated Use of AHP, GIS and RS.” ISPRS International Journal of Geo-Information, 2020.

De la Cruz, G., et al. “Present Variability and Future Change in Onset and Cessation of the Rainy Season Over Peru.” International Journal of Climatology, 2024.

Altea, L. “Perceptions of Climate Change and Its Impacts: A Comparison between Farmers and Institutions in the Amazonas Region of Peru.” Climate and Development, 2019.

Bush, M. B., N. A. S. Mosblech, and W. Church. “Climate Change and the Agricultural History of a Mid-Elevation Andean Montane Forest.” The Holocene, 2015.

Sales, R. A., et al. “Wet and Dry Events Influenced Colonization of a Mid-Elevation Andean Forest.” Quaternary Science Reviews, 2024.

Aybar, C., et al. “Construction of a High-Resolution Gridded Rainfall Dataset for Peru from 1981 to the Present Day.” Hydrological Sciences Journal, 2019.

Condom, T., et al. “Climatological and Hydrological Observations for the South American Andes: In Situ Stations, Satellite, and Reanalysis Data Sets.” Frontiers in Earth Science, 2020.

Gomez-Peralta, D., et al. “Rainfall and Cloud-Water Interception in Tropical Montane Forests in the Eastern Andes of Central Peru.” Forest Ecology and Management, 2008.

Rapp, J., and M. Silman. “Diurnal, Seasonal, and Altitudinal Trends in Microclimate across a Tropical Montane Cloud Forest.” Climate Research, 2012.

Potter, E. R., et al. “A Future of Extreme Precipitation and Droughts in the Peruvian Andes.” npj Climate and Atmospheric Science, 2023.