Diversified coffee agroforests modify microclimate conditions in comparison to monocultures, impacting the success of significant plant pathogens, such as Hemileia vastatrix, which causes coffee leaf rust (CLR). However, research is often limited to the dichotomous analysis of shaded agroforestry systems or unshaded monocultures, often overlooking the nuanced effect of shade tree trait diversity. Our study aims to determine the cumulative effects of shade tree canopy architectural characteristics and leaf functional traits in biodiverse agroforests on microclimate modifications and CLR incidence. We measured plot-level microclimate conditions (air temperature, relative humidity, leaf wetness duration, throughfall kinetic energy) in three single-stratum and two double-strata shade tree canopy treatments, including Erythrina poeppigiana, Terminalia amazonia, and Chloroleucon eurycyclum. Commonly reported canopy characteristics and leaf traits were compared to average microclimate conditions and CLR incidence levels. We found that shade tree trait expression significantly explained most microclimate conditions, and that two key shade tree traits (canopy openness, leaf area) significantly explain CLR incidence levels (R²  = 0.211, p = 0.036). Our results highlight the differences in microclimate conditions and CLR incidence among biodiverse agroforests, as well as the important explanatory power of shade tree traits. Specific effects of shade tree traits on pathogen dynamics can cirectly inform agroforestry system design (i.e. shade tree species selection) and sustainable coffee farm management practices (i.e. pruning practices).

Thirty circular plots were selected at the international coffee agroforestry research trial by the Centro Agronómico Tropical de Investigación y Enseñanza (CATIE), in Turrialba, Costa Rica located at 09°53´44” N and 83°40´7” W, at 685 m a.s.l. Measurements took place from May to July 2017. Within the site, Coffea arabica cv. Caturra (herein referred to as coffee) is planted in agroforestry systems under various management schemes. We included five distinct types of shade tree treatments in our study, including three single-stratum and two double-strata canopies: Erythrina poeppigiana (E), Terminalia amazonia (T), Chloroleucon eurycyclum (C), E. poeppigiana + T. amazonia (ET) and E. poeppigiana + C. eurycyclum (EC). All shade tree treatments were repeated in two amendment regimes: moderate conventional (MC) and intensive organic (IO). All shade tree and amendment treatment combinations were repeated in three distinct blocks within the CATIE farm. Plots were confined to the area directly beneath a single representative shade tree in the single-stratum canopy treatments (E, T, and C), or beneath the overlapping of both shade tree canopies in the double-strata canopy treatments (ET, EC). 

Microclimate measurements

Microclimate conditions were measured in five plots (each of the shade tree treatments) in the same block simultaneously for seven consecutive days, rotating between each of the amendment regimes and blocks, for a total of nine weeks between early May and late July 2017 (n = 63 days per treatment). Microclimate measurements were recorded every five minutes. We used HOBO® datalogger (Onset Computer Corporation, Bourne, MA) and HOBO® microclimate sensors (air temperature and relative humidity sensor (S-THB-M008) and leaf wetness sensor (S-LWA-M003)), which were installed next to the selected central shade tree between the main stem and canopy’s outer edge between the coffee plant rows. All air temperature and relative humidity sensors were installed on a stable vertical post and positioned above the average coffee plant height at about 2 m. Daily air temperature and relative humidity values were calculated based on data collected from midnight CST (or time of microclimate station installation) until 11:55 pm CST (or time of microclimate station removal). Start and stop times were kept consistent across all sampling plots. Leaf wetness sensors were positioned on the same vertical post at mid-coffee plant height (about 1 m), between the coffee plant row directed either north or south, laying horizontally and tilted at a 45^o angle facing east. Leaf wetness duration was calculated as the amount of time (hours) that the leaf wetness sensor was wet between 6:00 am CST and 7:00 pm CST, as all wetting and drying processes were captured within this timeframe. 

Total throughfall kinetic engery (TKE) of rain events was measured using Tübingen splash cups (Scholten, Geißler, Goc, Kühn, & Wiegand, 2011) with sand calibrated to 150-250 µm (Avelino et al., 2020), where total sand lost after each individual rain event was converted to the total TKE (J/m²) of the rain event using the equation developed by Scholten et al. (2011). Three splash cups were installed within each study plot: close to the central tree stem, mid-way between the central tree stem and canopy edge, and near the central tree’s canopy edge. Each cup was positioned on a stable vertical post at height of about 2 m (above the average coffee plant height). Measurements were collected in five shade tree plots (each of the shade tree treatments) in the same block simultaneously, rotating between each of the amendment regimes and blocks, for a total of two or four events, for a total of 18 days of TKE measurements per treatment. All measurements were collected simultaneously across the shade tree treatments on each sampling day.

Shade tree traits

For each shade tree, we measured total canopy height (m), canopy base height (CBH; m), canopy diameter (m) and canopy openness (%). In plots with one shade tree, the central shade tree was used for all measurements. In double-strata canopy plots, total canopy height was recorded as the tallest crown height, CBH was recorded as the shortest tree’s CBH, and canopy diameter was recorded as the largest crown diameter. Canopy openness was captured using hemispherical photography and analyzed using Gap Light Analyzer (Simon Fraser University, 1999). Hemispherical photographs were collected at a height above the average coffee plant height (about 2 m) at three positions under the central shade tree: close to the tree stem, in the middle of the canopy, at the edge of the canopy. The average canopy openness value of the three photographs, determined from the zenith angles 0^o to 45^o (Park & Cameron, 2008), were used in subsequent analyses.

We used a subset of six representative trees per species to measure leaf functional traits. Following protocols outlined by Pérez-Harguindeguy et al. (2013) and Pisek, Ryu, & Alikas (2011), five replicate leaf samples were collected from the lower canopy of each tree (n = 30 leaves per shade tree species), which were used to determine leaf area (cm²), specific leaf area (SLA; mg/mm²), leaf dry matter content (LDMC; mg/g) and leaf angle. The large leaflets of T. amazonia and E. poeppigiana were analyzed as separate leaves. In plots with double-strata canopies, final leaf trait values were calculated using the relative abundance-weighted mean of the shade tree species present in the plot, similar to Geißler et al. (2013), calculated based on the percent canopy area occupied by each shade tree canopy within the given plot.

Coffee leaf rust incidence

Within the boundaries each plot, six coffee plants of similar age were selected. CLR incidence was measured in July 2017, where coffee plants were assessed on the same day in each of the treatments. CLR incidence was estimated as the average percentage of leaves with CLR chlorotic spots, both with and without emerging mature uredospores, from three branches of varying height (Avelino, Romero-Gurdián, Cruz-Cuellar, & Declerck, 2012). 

Plot traits

Amendment: One of two amendment regimes used in this study, where MC is moderate conventional and IO is intensive organic.

ShadeTree: One of five shade tree treatments, including Erythrina poeppigiana (E), Terminalia amazonia (T), Chloroleucon eurycyclum (C), E. poeppigiana + T. amazonia (ET) and E. poeppigiana + C. eurycyclum (EC).

Block: Each shade tree treatment and amendment regime was repeated in three distinct blocks at the CATIE farm. 

TreeHeight: Total canopy height, measured in meters.

CBH: Canopy base height, measured in meters.

Diameter: Canopy diameter, measured in meters.

Openness: Canopy openness, measured as a percentage.

LeafAngle: Mean shade tree leaf angle, measured in degrees.

LA: Mean shade tree leaf area, measured in cm2.

SLA: Mean specific leaf area, measured in mg/mm2.

LDMC: Mean leaf dry matter content, measured in mg/g.

CLR_incidence: Plot-average CLR incidence at the plant-level, based on 6 coffee plants, measured as a percentage. 

T_max, T_min, T_range: Mean daily maximum, minimum, and range of air temperature per plot, measured as degrees Celsius.

RH_max, RH_min, RH_range: Mean daily maximum, minimum, and range of relative humidity per plot, measured as a percentage.

LW: Mean leaf wetness duration per plot, measured in hours.

TKE_max, TKE_min, TKE_range: Mean daily maximum, minimum, and range of throughfall kinetic energy per plot, measured as joules per square meter.

Intensity: Mean daily rainfall intensity, calculated as the amount of rainfall (mm) divided by the duration of the rainfall event (hours). 

Leaf traits

LeafRep: Five replicate leaf samples were collected for each shade tree species.

LeafAngle: Shade tree leaf angle, measured in degrees.

LA: Shade tree leaf area, measured in cm2.

SLA: Specific leaf area, measured in mg/mm2.

LDMC: Leaf dry matter content, measured in mg/g.

CLR

CoffeeRep: Six coffee plant replicates were selected for each plot.

CLR_incidence: CLR incidence at the plant-level, measured as a percentage.