A blueberry orchard that used to accumulate 500 chilling hours a year will accumulate 167 by mid-century, 33.4% of what it received, and the number of days on which it can accumulate them drops from 69 to 32. This is not a yield projection or a suitability scenario; it is the input the plant needs to come out of dormancy and open its buds, reduced to one third. In the same exercise, chill is reduced but frosts do not disappear: they keep falling until mid-March, with only a slightly lower probability and eight days later.
That is the kind of damage that does not show up in a conversation about climate and agriculture, because the conversation stays focused on how much will be produced. Temperature does not enter through yield; it enters through the process it governs, and each crop has a different one exposed. In coffee, it governs flowering synchrony; in Hass avocado, the coincidence of floral opening; in citrus, the separation between internal ripening and peel color; and in blueberry, chill accumulation. Two degrees more attack four different points and produce four losses that do not look alike.
Thermal time helps us know how much is left from the moment a process started. What decides whether that process starts well, or starts several times, is something else, and in the four crops we are going to talk about it is a different climatic variable. It is worth identifying which one is yours before putting instruments on the farm.
What a perennial cannot do is change varieties every year
The response of plants to high temperatures is well documented in annual crops such as barley, corn, rice, wheat, beans, peas, and soybeans, and information on the effect of elevated temperatures in perennial fruit crops is comparatively limited. That asymmetry in the literature reproduces the asymmetry in the field: whoever plants an annual reviews the decision every cycle, and whoever established a perennial committed for two decades to one genetic material, one elevation, and one exposure.
Added to that, a perennial lives by alternating growth and reproduction over several seasons, and to resume growth after an unfavorable period it needs specific temperature and photoperiod signals. An annual responds to the temperature of its single cycle. A perennial responds to the temperature of this cycle and the previous one, because the bud that will open now formed earlier and has been waiting for a signal. When that signal degrades, the damage is not corrected with the next planting.
Coffee: heat does not take away kilos, it takes away order
In three locations at 1,100, 1,400, and 1,900 meters, with average temperatures of 23.2, 21.4, and 17.1 degrees, flowering in a year with sufficient moisture was distributed predictably: at the lower site, peaks above 15% fell in December, with 23%, and January, with 27%; in the middle, it was evenly distributed across the first four months of the year, with 15 to 16% each month; higher up, it was later, with 22% in March and 27% in April.
What broke that order was not heat but the rainfall regime. After a dry July and August, September rain suddenly opened latent buds and produced a single peak of 79% at 1,900 meters, and 78% at 1,100 meters after 10.1 millimeters fell eight days earlier. The following year, with marked dry periods, flowering concentrated at 70% in January at the lower site, and at 54% and 84% at the other two elevations when February rains arrived. And when soil water was continuously sufficient, flowering was spread across three distinct periods in two of the three elevations. Intense dry periods concentrate flowering and steady rain distributes it over time; when there are no well-defined dry periods, coffee flowers all year, fruit ripening is not uniform, and harvest is therefore spread out.
That is where the loss lies, and it is not measured in kilos. A plot with one flowering event delivers its harvest in two or three passes. The same plot flowering in four waves delivers the same kilos spread over six, with green and ripe fruit on the same branch and a crew that has to be called out more times to pick less each time. The fruit is also not the same: the main and secondary harvests differ in the degree days the cycle needs, in the number of fruits per node, in firmness, diameter, fresh mass, and dry mass of the pericarp and bean. And the number of fruits per node explains much of those physical differences, because the space available for individual growth is reduced when node load is high.
Among those same studies appears a result worth keeping in mind. When the effect of variety on flowering synchrony and its variability was tested, there was no evidence of an effect; site and semester were the only factors with an impact. And among progenies of the same variety there were no significant differences in any of the fruit characteristics measured. As we already covered in our blog El niño 2026, lo que los productores agrícolas y pecuarios deben saber para anticiparse desde ya, which you can read here, the transition between Pacific phases is the event that most shapes yearly planning; in coffee, the effect comes through the rainfall regime and is reflected in the harvest calendar, not in the material’s potential.
Hass avocado, a narrow band where flowering synchronizes
Under Colombian Andean tropical conditions, thermal conditions are the main influence on Hass phenology, and a thermal and water gradient linked to altitude, between about 2,050 and 2,350 meters, determines how synchronously the crop processes develop. Anthesis is the stage most sensitive to the environment, although it remains comparatively stable near 2,200 meters, and degree days work as a reasonable predictor to define the altitudinal band where synchrony and productivity coincide.
That three-hundred-meter band is the operational data point. In Hass, the pollination window depends on female-phase and male-phase flowers being open at the same time, and that coincidence lasts hours within days. When the most sensitive stage of the cycle is precisely floral opening, losing synchrony does not spread the harvest into more passes like in coffee; it reduces it, because the flower that opened alone does not set fruit. A disordered coffee plantation produces the same and costs more to harvest. A disordered Hass orchard produces less.
The practical reading is that in Hass, elevation stops being a general suitability variable and becomes a synchrony variable. Two plots on the same farm separated by three hundred meters are not the same crop with shifted dates; one is inside the band and the other is outside it, and the one outside will set fruit erratically for reasons that no soil analysis will explain.
Knowing where each of your plots is losing
Do you know which stage of your crop decides the harvest, and what temperature your plot had during those two weeks? Most farms have records of field work and harvest, but they do not have the plot climate linked to the window when the crop was vulnerable, so at the end of the year there is a bad result and no way to attribute it. Harvis helps you define the critical window for each crop and each plot, associate each plot with its temperature and rainfall source, mark in the record what happened within that window, and compare seasons to see whether the bad year was climate or management. Write to us and we will set it up on the plots you already have defined.
Citrus and blueberry, two opposite ways to fail with the same extra degree
In citrus, heat separates two things the market demands together. Internal ripening and peel color respond to different thermal signals, and natural degreening of the rind is modulated by low temperature, with a thermal threshold for its onset. Under sustained warm conditions, the fruit reaches internal maturity while the peel remains green, so the lot is ready for consumption and rejected for appearance.
The effect on the pulp, moreover, does not go in one direction. Following fruit for six years in two dry-climate sites, in the arid site heat episodes above 35 degrees, especially at the beginning of summer, were associated with higher acidity and with less juice, fewer soluble solids, and less sugar; in the warm Mediterranean site, days between 27 and 33 degrees at the end of summer and beginning of autumn were associated with lower citric acidity, and sugar was independent of the microclimate. The same crop, two sites, two directions. The authors themselves note that the direct effect of temperature on internal ripening still needs confirmation, and that caution is part of the data.
In blueberry, the point of attack comes before all that. The projected chill accumulation by mid-century starts four days later and ends 28 days earlier, the total falls from 500 to 167 hours, and the days with accumulation drop from 69 to 32. The chill portions calculated with the dynamic model are cut almost in half. Lack of adequate chill affects bud break and flowering time, which can desynchronize flowering from pollinators, cause floral bud death, and impact yield; and an early exit from dormancy leaves the plant more exposed to spring low temperatures. In other words, warmer winters and at the same time a higher risk of cold damage.
That is where the question we set aside gets answered. In the same chamber, with the same warming applied, two blueberry cultivars responded differently in bud break, early flowering, and full flowering, with significant cultivar-by-treatment interactions; one did not complete early flowering under elevated temperature and the other did not complete full flowering under ambient temperature. In blueberry, genetic material is a real lever, and that is why specific low-chill breeding exists. In coffee, genotype did not explain the flowering pattern or the fruit characteristics. Changing varieties works in one crop and not in the other, and confusing them costs a full plot renewal.
What these figures cannot tell you about your farm
The first limit is direction. In three of the four cases, the effect of heat is mediated by another variable: rainfall in coffee, altitude in Hass, accumulated chill in blueberry, and only in citrus is there a direct relationship between temperature and fruit quality, which also changes sign between sites. Attributing a single effect in one direction to heat is the mistake these figures help avoid.
The second limit is the origin of the data. The blueberry projections come from a high-emissions climate scenario applied to a specific producing region, and the coffee series correspond to two years with a warm event in one and a cold episode in the other. Neither describes your farm, and before comparing your plot with them it is worth knowing where your own data come from. Correcting the temperature of another station by elevation requires two gradients, not one, because in Colombia’s coffee zone the annual mean maximum falls 0.7 degrees per 100 meters and the minimum only 0.5. A sensor without a radiation shield deviates by about one degree in full sun, and gridded data underestimate the maximum and overestimate the minimum, compressing the diurnal range.
The third limit is that none of these studies measured a historical shift in dates in Latin America. There are measured relationships between climatic variables and crop processes, and there are projections for other producing regions, but there is no long published series saying how many days flowering of a Latin American crop shifted per decade. Whoever offers you that figure is building it with a model, not reading it from a record.
The fourth point is about transferring the figures. The 500 and 167 chilling hours, the 2,050 to 2,350 meter band, and the thresholds of 35 and 27 to 33 degrees correspond to specific crops, sites, and scenarios. What transfers is the procedure: identify which process governs the harvest in your crop, find the time window in which that process occurs, and measure temperature within that window on your plot. The threshold table does not transfer.
Fourteen days, a thermometer, and the record of what came out second grade
The exercise consists of measuring one window instead of the whole year. First define which process decides your harvest: flowering synchrony in coffee, anthesis in Hass, peel color change in citrus, bud break in blueberry. That process occupies a short stretch of the calendar and is the only stretch you really need to instrument this year.
Set a fourteen-day window around that process, seven days before and seven days after the date it happened last year. Place a maximum-minimum thermometer with a radiation shield inside the plot, at the height of the middle third of the canopy, and record maximum and minimum every day during those fourteen days. If your crop is blueberry, the window is not fourteen days but the whole winter, and what you count are not degree days but hours below the chill threshold of your cultivar.
During those same fourteen days, write down what you see, one line per day. In coffee, how many trees in the plot are flowering. In Hass, whether there are open flowers and of what type. In citrus, whether the peel has started to turn. Fourteen lines per season, one per day, is all the record this part requires.
Close the count with the result. When you sell, note what proportion of the plot came out second grade and for what reason: small fruit, insufficient color, uneven ripening, or how many harvest passes were needed. That is the number you will later be able to compare against the temperature in the window, and it is what turns the exercise into an account rather than a weather log.
Repeat it in a second plot that is at least one hundred meters higher or lower. Two plots on the same farm in the same season share regional climate and differ in local climate, which is exactly the separation you need to know whether the problem is the year or the site.
Each crop fails where it is narrowest
Warming is not a factor that pushes all crops in the same direction with different intensity. It is pressure that finds, in each species, the process with the least margin and breaks it first. In coffee, that process is the synchrony of a flowering event that depends on the contrast between dry weather and rain, and when that contrast disappears the harvest still comes, but it comes spread out and more expensive to pick. In Hass, it is the coincidence of a floral opening that can only be sustained within a narrow elevation band. In citrus, it is the coupling between internal maturity and external color, which respond to different signals. In blueberry, it is an invisible input that accumulates in winter and is already being lost by half. A producer with four crops has four different problems, and the technical response that works in one does not work in another.
What you can do starting Monday fits into five decisions. Write in one line which process decides the harvest of your main crop, because that name determines what you will measure and what you can stop measuring. Define the fourteen-day window in which that process occurred last year and place a thermometer with a shield inside the plot for this year’s window, because measuring the whole year is what makes nobody measure anything. Record in fourteen lines what you see in the plot during that window, since climate without observation explains nothing. When you sell, record what proportion came out second grade and why, because without the result the window is a datum without consequence. And before deciding on a renewal because of climate, verify whether in your crop the material is a real lever, because in blueberry cultivars respond differently to the same warming and in coffee genotype did not explain the pattern. The producer who ends the season with fourteen days of temperature, fourteen lines of observation, and a second-grade percentage will know whether the bad year was climate or management; the one who does not have them will blame climate anyway, which is what happens when there is nothing else to decide with.
Do you want the critical window of each plot, its temperature, and the sales result in the same record, so you can know where each crop is losing? Write to us on WhatsApp or schedule a demo and we will show you how Harvis links plot climate with what ultimately came out of it.
References consulted
- Impacts of projected elevated temperature on estimated chilling accumulation and bud break of rabbiteye blueberries. (2025). Preprint, bioRxiv. https://doi.org/10.1101/2025.07.08.663765
- Vélez-Arango, B. E., Jaramillo-Robledo, Á., Chaves-Córdoba, B. and Franco, M. (2000). Distribution of coffee flowering and harvest at three elevations. Avances Técnicos Cenicafé, 272, 1–4. https://doi.org/10.38141/10779/0272
- Unigarro, C. A., Imbachi, L. C., León-Burgos, A. F., Cayón-Salinas, D. G. and Flórez-Ramos, C. P. Harvest-Dependent Variability in the Ripening and Fruit Characteristics of the Cenicafé 1 Variety of Coffee. Crops, MDPI, 6(4), 75. https://doi.org/10.3390/crops6040075
- Quantification and Qualification of Floral Patterns of Coffea arabica L. in Colombia. (2023). Plants, MDPI, 12(18), 3332. https://doi.org/10.3390/plants12183332
- Rebolledo-Roa, A. and collaborators. (2025). Climatic Drivers of Flowering Synchrony in ‘Hass’ Avocado Under Tropical Andean Conditions. Plants, MDPI, 14(24), 3822. https://doi.org/10.3390/plants14243822
- Influence of temperature on the beginning of degreening in lemon peel. Scientia Horticulturae. https://www.sciencedirect.com/science/article/pii/S0304423812003482
- Mesejo, C. and collaborators. (2024). Examining the impact of dry climates temperature on citrus fruit internal ripening. Scientia Horticulturae, 337, 113501. https://doi.org/10.1016/j.scienta.2024.113501
- Spatial and temporal variability of air temperature in the Colombian coffee zone. Investigaciones Geográficas, Boletín del Instituto de Geografía. https://doi.org/10.14350/rig.38707
- Holden, Z. A., Klene, A. E., Keefe, R. F. and Moisen, G. G. (2013). Design and evaluation of an inexpensive radiation shield for monitoring surface air temperatures. Agricultural and Forest Meteorology, 180, 281–286. https://research.fs.usda.gov/treesearch/44116
- Evaluation of Daily Temperature Extremes in the ECMWF Operational Weather Forecasts and ERA5 Reanalysis. (2024). Atmosphere, MDPI, 15(1), 93. https://doi.org/10.3390/atmos15010093
- Climate changes altered the citrus fruit quality, a 9-year case study in China. (2024). Science of the Total Environment. https://www.sciencedirect.com/science/article/abs/pii/S004896972401547X
- Examining the Role of Low Temperature in Satsuma Mandarin Fruit Peel Degreening via Comparative Physiological and Transcriptomic Analysis. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9328020/
- Mesejo, C. and collaborators. (2012). Relationship between soil temperature and fruit colour development of ‘Clemenpons’ Clementine mandarin. Journal of the Science of Food and Agriculture. https://doi.org/10.1002/jsfa.4600
- Flowering and Fruiting of Coffea arabica L., A Comprehensive Perspective from Phenology. (2025). Plants, MDPI, 14(21), 3396. https://doi.org/10.3390/plants14213396
- Chill topped historical Arabica coffee yield loss among climate stressors in Yunnan, China, followed by drought. (2025). npj Sustainable Agriculture. https://doi.org/10.1038/s44304-025-00092-5
- Influence of Excessively High Temperatures on the Fruit Growth and Physicochemical Properties of Shiranuhi Mandarin in Plastic-Film Greenhouse Cultivation. Plants, MDPI, 10(8), 1525. https://doi.org/10.3390/plants10081525
- Food and Agriculture Organization of the United Nations. Ripe for change, adapting avocado production to a changing climate. https://openknowledge.fao.org/server/api/core/bitstreams/c1062d72-ae62-43b1-97ae-aab32c75cadf/content