3.1 Microclimate
The monthly minimum temperatures varied according to the height of measurement, it was observed lowest values closer to the ground surface (0.5 m), increasing gradually with height (Fig. 1).
This happened due to the thermal inversion, in which on cold autumn/winter nights, in conditions of atmospheric stability, without clouds, low air humidity and absence of winds, intense radiative losses from the surfaces occur. As the cold air is denser and with continuous cooling, temperatures remain lower in the air layers closer to the soil surface (Caramori et al., 1999; Morais et al., 2006).
The temperature measurement points closer to the canopy showed greater differences between treatments, especially in the coldest months (May to August). However, at the height of 2.0 m the differences between treatments were lesser. This may have occurred due to the limited size of the plots, which may have diluted the effect of the shading of the trees with the penetration of external air inside the evaluated plots. Furthermore, the measurement height of 2.0 m may have been too high to assess the environment inside the plantation, being more subject to interference from outside air. In this sense, the temperature of the leaves and at 0.5 m better represent the cooling that occurs inside the canopy.
The minimum leaf and air temperatures at 0.5 m in the autumn/winter (May-August) months of 2020 were higher in coffee under AFS, compared to full sun, indicating the potential of this cultivation practice to protect plants under conditions of sub-optimal temperatures (Fig. 1). Trees provide this effect because on cold nights, when the surface progressively loses heat, mainly influenced by radiation and conduction processes, their crowns intercept the radiation emitted by the surface, maintaining a warmer environment.
During an episode of weak frost, on August 21 and 22, 2020, it is observed during the cold dawn, that the air temperatures at 0.5 m of the coffee trees cultivated at full sun were lower than those of the coffee trees under AFS, mainly from 00:15 (Fig. 2a).
Comparing the AFS, in coffee trees under the species C. floribundus, the highest temperatures were recorded, being this the species that presented the densest canopy (63%). A study during frost nights showed that coffee leaf minimum temperatures were higher when cultivated in a bracatinga (Mimosa scabrella) system than in open-air coffee plants (Caramori et al., 1996). In our study, coffee trees cultivated under C. floribundus reached a temperature of 3.9 ºC higher than those cultivated under full sun. These results confirm that tree species have an effect of retaining heat inside the canopy, preventing the drop in temperatures and protecting coffee trees in extreme low temperature events. It should be noted that, even if infrequent, just one frost event, depending on the intensity of the phenomenon and the age of the plant, is enough to damage coffee plants, reduce productivity or even cause plant death (Caramori et al., 1996).
Regarding maximum temperatures, a greater effect was observed when temperatures were higher during the spring (Sep-Dec, 2020) and summer (December-March) seasons (Fig. 2b). It was found that the species S. macranthera and C. floribundus, with leafy canopies and high levels of shading, attenuated the maximum temperatures throughout the period, providing a more suitable environment for coffee cultivation. On the other hand, coffee trees cultivated in consortium with the species M. oleifera presented the highest values of maximum temperatures, even higher than the treatment full sun, in most of the evaluated period. This occurred because this species remained practically leafless, with only 7% of solar radiation intercepted, also associated with the limited and reduced size of the plots, which cause heat transport by convection from the adjacent plots.
The temperatures of coffee leaves on October, 7, 2020 were extremely high during the day, reaching 46 ºC at full sun and shaded with M. oleifera (Fig. 2c). In the AFS with C. floribundus and S. macranthera, softened the temperatures of the coffee leaves up to 7.9 ºC and 7.7 ºC, respectively. Coffee plants do not tolerate high temperatures for prolonged periods, which can cause damage to the leaves and photosynthetic apparatus and, if they occur in the flowering phase, it causes abortion of flower buds and poor flower formation.
The minimum soil temperatures under coffee trees grown in full sun recorded the lowest averages in the cold months, showing a greater thermal amplitude in soil temperatures (Fig. 3a).
Soils under AFS retained more heat, especially in July 2019, when soils under M. oleifera, C. floribundus and S. macranthera had averages of 14.4°C, 15.3°C and 16.6°C, respectively, while the coffee trees in full sun registered an average maximum temperature of 9.1°C. Note that the shading density did not affect soil temperature, since there were no pronounced differences between the shading species.
The soil of coffee trees cultivated at full sun showed the highest values of maximum temperature and highest thermal amplitude during almost the entire period of the evaluation (Fig. 3b). Comparing the coffee plants under AFS, it is noted that the values of maximum soil temperature were proportional to the shading density, thus, the more shaded the coffee plants, the lower the maximum temperatures. The soil under shaded coffee trees had lower average maximum temperatures throughout the period, with values of 20.4°C for C. floribundus, followed by the species S. macranthera with 21.5°C and M. oleifera with 23.3°C. Soils under coffee trees in full sun reached an average maximum temperature of 29.3°C. The large difference in maximum soil temperatures of coffee trees in full sun and under AFS shows the important role of trees in mitigating the coffee microclimate within the soil profile.
The global warming forecast issued by the Intergovernmental Panel on Climate Change (IPCC, 2014) has caused great concern for the coffee agribusiness. The use of afforestation would be one of the agricultural techniques of mitigation for the possible scenario of global warming as highlighted in some studies. For instance, in coffee crops intercropped with pigeon pea (Cajanus cajan), leucena (Leucaena leucocephala), gliricidia (Gliricidia sepium), and macadamia nuts (Macadamia tetraphylla) (Coltri et al., 2019), with rubber trees (Hevea brasiliensis) (Assad et al., 2020; Zaro et al., 2020), which indicates the potential of this agricultural practice to reduce the effects of global warming on the crop to mitigate high temperatures.
Overall, there was a variability in relative humidity throughout the period, which is associated with the rainfall index, number of days with rain, winds, cloud cover and air temperature (Fig. 4a).
Overall, there was an increasing in the relative humidity of the air as the rainfall index increases. It is noted that throughout the period there were no great variations between treatments, with values of 44% − 45%. Regarding soil moisture, it is observed that in August, 2019 the soils with M. oleifera species stood out with greater moisture than the others at both distances. In January, 2020, regardless of the distances (0.7 and 1.0 m) of coffee tree trunk sampling, there was no difference between treatments (Table 2).
In August, 2019, the soil was in water deficit, showing that under intercropped coffee trees the M. oleifera species had the potential to retain moisture on the soil surface in periods of drought (Fig. 4b). In contrast, at the beginning of January, 2020, the soil was in excess of water and showed no differences between treatments. Studies with arabica coffee (Morais et al., 2006) and conilon coffee (Guimarães et al., 2014) showed that under AFS there is higher soil water content than in full sun crops.
3.2 Interaction between PAR and coffee bean yields
The species C. floribundus and S. macranthera with 63% and 58% of shading, respectively, presented higher mean values than M. oleifera (7%), due to the denser crowns, with leaves present throughout the year, which vigorously intercepted solar radiation by reflection and absorption, while the latter species had few leaves in the canopy throughout the experimental period. In all treatments, the photosynthetically active radiation (PAR) increased as the months progressed from winter to spring, due to terrestrial declination (Data not shown). For each species, there was variability in PAR in the different months due to changes in leaf area and canopy density. There were no major differences in PAR between coffee trees at full sun and under M. oleifera, this was maybe due to the small amount of leaves observed in this species, unlike the other trees (Fig. 5a).
In 2017/2018 growing season, coffee beans grown in full sun had the highest yield (3576 kg ha− 1), while in the AFS with M. oleifera (3006 kg ha− 1) and C. floribundus (2128 kg ha− 1) had intermediate yields without significant differences, however when with S. macranthera had the lowest production (1881 kg ha− 1) (Fig. 5b). There is a high positive correlation between solar radiation and coffee productivity, which suggests that the amount of radiation that hit coffee trees under denser shading indirectly decreased grain production (Fig. 5c). From an agronomic point of view, AFS with S. macranthera and C. floribundus should be managed with greater spacing between the trees and/or more pruning to have greater solar radiation in the coffee trees.
In Mexico, shading of coffee plants up to a limit of 50% did not reduce coffee productivity (Soto-Pinto et al., 2001). For small Peruvian producers, Arabica coffee in the shading system with trees shows no negative relationship with economic performance, possibly due to revenue from other products, such as timber, which provides these farmers with an extra source of income (Jezeer et al., 2018). For both fertilizer and weed suppression functions, Staver et al. (2001) suggested that shade tree stratum in coffee plantations should have at least two tree species of contrasting phenology to maintain shade levels within the desirable range of 35 to 65% for most coffee regions. Shading of Arabica coffee trees combined with altitude might improve bean quality (Tolessa et al., 2017). Coffee trees in shaded cropping systems with attenuation of approximately 20% of global solar radiation did not produce negative effects on coffee production (Caramori et al., 1996; Siles et al., 2010). It is worthwhile to note that in India, there was a positive effect of shade tree diversity on both coffee (Coffea canephora) production and quality (Nesper et al., 2017).
3.3 Soil chemical and microbial indicators
Coffee cultivation in full sun had the lowest concentrations soil carbon and the most acidic pH when compared to shaded crops with trees, while under shaded with C. floribundus and M. oleifera showed the highest values of Ca2+ and Mg2+ and base saturation, indicating that they are having an efficient management (Table 1).
Table 1
Soil chemical properties in the 0-0.10 m, 0.10–0.20 m and 0.20–0.40 m depth under coffee trees using agroforestry systems (AFS) and full sun.
Treatments
|
apH
|
Corg
|
P
|
K
|
H + Al
|
Ca
|
Mg
|
BS
|
|
_g kg̶1_
|
_mg kg̶1_
|
___________Cmolc kg̶1 soil___________
|
__%__
|
|
0-0.10 m depth
|
Full sun
|
5.45
|
18.65
|
14,65
|
0.90
|
4.45
|
5.15
|
2.55
|
65.67
|
M. oleifera
|
5.83
|
18.75
|
9.00
|
0.82
|
3.76
|
6.27
|
3.20
|
73.10
|
C. floribundus
|
5.70
|
19.91
|
12.80
|
0.88
|
4.22
|
6.30
|
3.22
|
70.85
|
S. macranthera
|
5.50
|
19.21
|
17.30
|
0.90
|
4.78
|
5.30
|
2.34
|
64.00
|
|
0.1–0.20 m depth
|
Full sun
|
5.18
|
18.50
|
11.23
|
0.87
|
5.17
|
4.44
|
2.04
|
58.60
|
M. oleifera
|
5.58
|
18.24
|
9.13
|
0.69
|
4.39
|
5.15
|
2.56
|
65.52
|
C. floribundus
|
5.33
|
19.04
|
9.35
|
0.73
|
4.54
|
4.50
|
2.27
|
62.12
|
S. macranthera
|
5.28
|
18.85
|
14.08
|
0.83
|
4.61
|
4.82
|
2.13
|
62.74
|
|
0.2–0.40 m depth
|
Full sun
|
4.90
|
14.80
|
6.73
|
0.71
|
5.46
|
3.48
|
1.59
|
51.38
|
M. oleifera
|
5.15
|
15.13
|
5.03
|
0.52
|
4.64
|
4.00
|
1.96
|
57.86
|
C. floribundus
|
5.00
|
15.61
|
4.55
|
0.59
|
5.29
|
3.65
|
1.79
|
53.21
|
S. macranthera
|
5.15
|
15.41
|
6.13
|
0.62
|
4.44
|
3.82
|
1.78
|
58.27
|
Note: pH in CaCl2 (0.01 mol L− 1); Corg (total organic carbon) by Walkley-Black method; P and K+ (HCl 0.05 mol L− 1 Mehlich-1; H + Al in SMP; Ca2+, Mg2+ and Al in KCl 1 mol L̶1; BS (base saturation). |
The soil under coffee cultivated with tree shading had higher concentrations of organic carbon (Corg), especially the species C. floribundus and S. macranthera, which increased 3% and 7% in relation to full sun. This result indicates the potential of AFS to increase Corg, which allows the supply of nutrients along the coffee tree with the mineralization of organic matter, in addition to maintaining soil moisture.
Overall, microbial biomass C (MBC) values from sampling in August, 2019were higher than those from January 2020. In August 2019, for MBC, there were no significant differences between treatments. In contrast, in January 2020, soil microbial carbon biomass was higher (p < 0.05) at 0.7 m than at 1.0 m (Table 2).
This was possibly due to the greater proximity of the root system and also to the greater self-shading of the coffee trees, whose milder temperature condition favoured microbial growth. In both season, winter (August, 2019) and summer (January, 2020), the microbial biomass of N was not altered by shading or by the soil sampling site (Table 2).
In previous study in this same clayey soil, the microbial carbon biomass was greater in the projection of the coffee tree canopy than in the inter-row of the plants (Andrade et al., 1995). Coffee trees cultivated in full sun showed lower amounts of microbial biomass C when compared to shade cultivation (Guimarães et al., 2017). However, for the coffee-araucaria agroforestry system it did not affect the distribution of the microbial groups, microbial biomass carbon, microbial activity and metabolic quotient in soil (Melloni et al., 2018). In the agroforestry system, the tree provided soil protection keeping the soil temperature more stable and at milder and more suitable levels for the growth of microorganisms in the soil.
Overall, basal soil respiration (BSR) and metabolic quotient (qCO2) at both sampling locations of (0.7 and 1.0 m) were higher in January 2020 than in August 2019 (Table 2). This increase might occur due to the higher temperature and rainfall rates that favour a higher phytomass in the system and consequently higher BSR and qCO2. In August 2019, for BSR, there was only a significant difference at location of 1.0 m away from the coffee tree trunk, in which the lowest value was in full sun. In January 2020, regardless of the sampling site (0.7 m or 1.0 m), there was no difference in BSR. Possibly, there was the influence of weeds in the coffee inter-row in all treatments which provides greater nutrient cycling through the addition of organic material, greater protection of the soil from solar radiation, increasing biological activity and soil resilience. In organic coffee cropping systems, soil microbial activity and qCO2 in summer were higher due to higher temperatures and precipitation (Pimentel et al., 2011).
In August 2019, for acid phosphatase activity, there was no difference with shading of the coffee tree. while in January 2020, at a distance of 1.0 m, this enzyme had higher activity in all shaded coffee crops than in the soil in full sun. In January, 2020, regardless of the coffee shading treatment, acid phosphatase activity showed lower rates in soil samples at 0.7 m than at 1.0 m from the coffee tree trunk. In August, 2019, the alkaline phosphatase activity had lower values at 0.7 m in the full sun and with C. floribundus (Table 2). Overall, coffee AFSs positively influenced soil microbial activity, improving soil characteristics, as the increase in organic residues favoured the diversity and functional microbial structure.
3.4 Soil chemical, microbial and microclimate correlation
In August, 2019, in the PCA correlation plane between chemical and microbial variables, treatments with C. floribundus and M. oleifera showed high values of the variables that most contributed to PC1. While in January, 2020 soil sampling, the first two PCs show the position of AFS with S. macranthera in the negative coordinate of the x and y axes, which indicates a lower correlation of the variables that contributed most to the PC2. Most of the microbial variables grouped in the quadrant in which it was placed the coffee AFS with C. floribundus (Fig. 6a).
In January, 2020, the hottest and rainy season, a greater correlation between the variables in the treatment with C. floribundus was also observed, with the acid phosphatase enzyme variables (0.7 m and 1.0 m) having a strong correlation with the Corg variable in a more isolated way, different from that observed in the dry season, in August, 2019 (Fig. 6b). Coffee shaded with C. floribundus have more favourable conditions for the development of soil microbial community possibly because they promote a greater contribution of organic material deposited on the soil, in addition to providing plant substrates in different degrees of decomposition.
In August, 2019 sampling, in the correlation between the microclimatic and microbial variables, the first two PCs of the analysis are presented, the M. oleifera and C. floribundus AFS presented high values of the variables that most contributed to the PC1. The systems in full sun and AFS with S. Macranthera showed low values of the variables of greater availability for PC1, that is, less interaction with air maximum temperature; basal soil respiration at 0.7 and 1.0 m; qCO2 at 0.7 and 1.0 m; air minimum temperature and acid phosphatase enzyme at 1.0 m(Fig. 6c).
The microbial variables most correlated with the 1st axis were, in descending order, qCO2 at 1.0 m; acid phosphatase enzyme at 0.7 m; basal soil respiration at 1.0 m; MBC at 1.0 m; MBN at 0.7 m and qCO2 at 0.7 m. In the second axis, the MBN at 1.0 m stood out. The microclimate variables that most correlated with the 1st axis were, in descending order, air maximum temperature, soil maximum temperature and air average temperature. For the 2nd axis, air maximum temperature at 2.0 m stood out. The position of the treatments, in full sun and with S. Macranthera, in the negative coordinate of the x and y axes indicates a lower correlation of the variables that most contributed to the PC2 obtained in the PCA. However, it can be observed that the variables air minimum temperature at 2.0 m, air average temperature at 2.0 m, soil moisture at 0.7 and 1.0 m, and MBN at 1.0 m were influenced by these treatments in the driest and coldest season, in August 2019 (Fig. 6c).
In January 2020, the position of the AFS with S. macranthera and C. floribundus in the negative coordinate of the x- and y-axes indicates a lower correlation of the variables that most contributed to the PC2 obtained in the analysis. Most of the microbial variables that correlated with the 1st axis was in descending order, acid phosphatase at 0.7 m; basal soil respiration at 0.7 m and 1.0 m; and MBC at 1.0 m. In the second axis, MBN at 0.7 m stood out. The microclimate variables that most correlated with the 1st axis were in descending order air average temperature at 2.0 m, soil minimum temperature, and coffee leaf minimum temperature. For the 2nd axis, air maximum temperature stood out (Fig. 6d). The interaction between microclimate and soil microbial community demonstrates that higher soil moisture rates and lower soil and air thermal amplitudes found in agroforestry systems can improve better soil quality.