Part I, Preliminary results
At the beginning of the current season, the Avium technical team published an article entitled “Preliminary analysis of the loss and sudden death of flower buds in cherry trees at the beginning of spring”, in which it was revealed that this phenomenon is multifactorial, mentioning various factors that could have been triggers, such as:
- Post-harvest stage.
- Recess entry.
- Dormancy-breaking strategies.
- Weather events.
Given this atypical season, with ups and downs from a production point of view, a more thorough study of the components of this equation must be carried out, where a useful analysis to identify the factors with the greatest influence on production potential is the principal component analysis (PCA).
In practice, this analysis assumes that each proposed variable represents a dimension, which makes macro analysis difficult. The PCA seeks to simplify the general visualization of the data, reducing the dimensionality to two main components that explain a significant percentage of the variability of all the data collected.
The Avium technical consulting and R&D team is jointly developing a critical analysis to obtain clear conclusions about the factors involved in the production decline that occurred during this season. Some preliminary results are presented below:
1. Multivariate analysis of the current season.
The analysis was initially carried out exclusively on the Santina and Lapins varieties, in nearly 70 productive blocks of different orchards with a planting year prior to 2019, which were separated into three different zones: early, medium and late.
To include categorical variables in the PCA, such as soil texture, agroclimatic zone, among others, the "one-hot" coding method was used, where new variables were created with binary values from 0 to 1 that represent each category of the original variable.
The factors that make up the ACP are:
- Agroclimatic zone (early, middle and late).
- Base chill hours 7.2 °C, 2023 season (accumulated from May 1 to the date of application of dormancy breakers or, failing that, until July 31).
- Precipitation for the 2023 season (mm accumulated from May 1 to September 30).
- Soil texture (sandy loam, loam and clay loam)
- Projected and/or actual yield in cases of finished crops (kg/ha).
It is under development to incorporate new variables that can have a great impact and thus correct the expected performance to the actual performance of the season.
It is a common denominator during this season that early agroclimatic zones suffered a greater impact on productive potential, directly related to the low accumulation of cold hours and degree days after sprouting. However, it is absolutely necessary to record the behavior of the season with real data and objective statistical analysis.
Figure 1 presents the PCA by individuals that explains 58% (PC1 36.1% + PC2 21.9%) of the total variability of the data, where the proximity between points (observations) refers to their similarity between them. Each point cloud represents a zone (symbology indicated in the legend), highlighting that the early zone presents a particular behaviour pattern, different from that of the middle and late zones.
Figure 1. Individual principal component analysis. Identification of homogeneous groups by agroclimatic zone. CP: principal component.
In the analysis carried out, it can be identified that the early zone presents a particular behavior during this season, however, what relationship do the variables included in the analysis have?
Figure 2 presents the PCA from a variables perspective. The legend (cosine squared - cos2) indicates the strength with which the variable is explained by the principal component, where close to 1 corresponds to a strong relationship and close to 0 to a weak one. The closeness between variables in terms of direction (angle) gives indications of trends or correlations between them, the latter of which must be verified by further analysis. Two groups of variables were identified and are enclosed in blue.
The expected performance (ER) was standardized into low, medium and high such that:
- RE < 5,000 kg/ha: “Low”
- 5,000 kg/ha <= RE <10,000 kg/ha: “Medium”
- RE >= 10,000 kg/ha: “High”
Figure 2. Principal component analysis. CP: principal component. RE: Expected yield; PP (mm): Precipitation (mm); HF: Base chill hours 7.2 ° C.
Some preliminary conclusions can be drawn from the figure above.
Cold accumulation, expressed in chilling hours (CH) and high expected yield (HE_High), share a similar projection, suggesting a positive correlation between them (blue circle on the right). On the other hand, clay loam soils have a strong correlation with low expected yield (HE_Low) highlighted in the blue circle on the left. When performing a correlation test for both cases, under statistical analysis, the p-value is less than 0.05, meaning they present a statistically significant correlation.
In summary, according to the initial analysis applied exclusively to the current season, a lower accumulation of cold hours and soils with a greater capacity to retain moisture were the factors that most influenced cases of low productivity in the crop.
2. Agroclimatic perspective.
An important parameter to quantify is the effective pollination period (PEP), for this a powerful agroclimatic indicator is the “match” between the number of hours with radiation above 300 W/m2 and temperatures above 15 °C.
As can be seen from Table 1, the number of hours with favourable conditions for pollination during the first week of September is extremely low. The hours compared to the previous season were 90% during the first week and 26% less during the second week of September.
Consider as a base parameter for this analysis, that 5 favorable hours for bee flight can represent 35 hours per week.
Table 1. Hours with radiation > 300 W/m2 and temperature > 15 °C. Season 2023-2024. VI and VII region, Chile.
An analysis of Table 2 reveals that the accumulation of base degree days of 4.5°C (base temperature as thermal accumulation to exit dormancy) during September remained stable, therefore, the limiting factor for bee flight and pollination was radiation during these dates.
Table 2. Weekly accumulation rate of base degree days 4.5 °C. Season 2023-2024. VI and VII region, Chile.
An interesting factor to observe is winter precipitation, where the analysis of several seasons (analysis in progress) showed, preliminarily, to have a negative and statistically significant correlation with the accumulation of cold hours and thermal accumulation. This raises a new challenge of analyzing the impact of precipitation at specific times on productive potential.
Figure 3 shows the weekly precipitation rate from May 1, 2023, to mid-November, where "peaks" corresponding to events that occurred from late August to mid-September are clearly identified, which is significantly detrimental to the proper phenological evolution of the cherry tree. In this season, it was observed that in early harvest areas, in the aforementioned period, the phenology was between exposed bunches and full bloom, which produced a very slow progress in the initial stages of the phenological states of the crop, where the adequate conditions for its correct development were not met.
Figure 3. Precipitation rate mm/week. Season 2023-2024. From 01.05.2023 to 16.11.2023.
4. Preliminary conclusions
With a more advanced season and more robust information data for analysis, the productive decline of the 2023-2024 season can be related to the following points:
1. Base chill hours 7.2 °C.
In the cases evaluated, the accumulation of cold is below what was expected and even below the historical average. This situation is very marked in early production areas.
2. Thermal accumulation (Degree days).
Once the plant receives the stimulus to break dormancy, the success of a good flowering and budding depends largely on the increase in heat gain, which this season could only be seen in mid and late production areas, mainly explained by the flowering date and the subsequent result of more favourable ambient temperatures. However, in early areas there was an irregular budding, delayed in the upper third of the tree, where, particularly, the largest share of final kilos is positioned.
3. Soil texture.
One of the factors that has been repeated in those cases of low production is the soil texture. The PCA indicates a direct relationship between low production and soil texture with greater moisture retention. Without a doubt, in those fields that have this type of soil, the cultural work of soil breaking, added to good drainage and rapid evacuation of water in case of flooding, helped in part to maintain the health of the plant, preventing it from collapsing due to root asphyxiation.
4. Precipitation.
In ongoing studies corresponding to the future Part II of this article, it was identified that precipitation during the dormancy period and after dormancy has an impact on the accumulation of cold hours and thermal accumulation.
5. General.
Low productivity scenarios occurred in those cases where the factors described above were perfectly combined. This situation can be clearly observed in areas where low cold accumulation, soil texture (heavy soils) and above-historical rainfall towards the end of winter accentuated low productivity.
Based on the information collected to date and supported by statistical analysis, it can be concluded, preliminarily, that the precipitations caused a scenario of cold accumulation at the limit of the requirement of the species and/or in some cases lower. At the same time, they maintained high levels of humidity in soils with a high retention capacity until spring, impacting soil temperature, accumulation of degree days and pollinator activity in spring.
Although this season's rainfall was well above the average of recent years, the phenological moment of the crop at the time of the event was probably the greatest impact, over and above the accumulated volume of rainfall.
