Contrasted agronomical and physiological responses of five Coffea arabica genotypes under soil water deficit in field conditions

Abstract

Introduction: Breeding programs have developed high-yielding Coffea arabica F1-hybrids as an adaptation against adverse conditions associated with climate change. However, theresponse to drought of coffee F1 hybrids has seldom been assessed.

Methods: A trial was established with five C. arabica genotypes (2 pure lines: Catimor and Marsellesa and 3 F1 hybrids: Starmaya, Centroamericano and Mundo Maya) planted under the leguminous tree species Leuceana leucocephala. Coffee growth, yield and physiological responses were assessed under a rain-fed (control: CON) and a rainfall reduction treatment (RR) for 2 years.

Results: The RR treatment created a long-term rainfall deficit in a region with suboptimal temperature similar to those predicted by climate change scenarios. Moreover, the RR treatment reduced soil water content by 14% over 2 successive years of production and increased hydric stress of the three F1-hybrids (leaf water potentials averaged -0.8 MPa under RR compared with -0.4 MPa under CON). Under RR, coffee yields were reduced from 16 to 75% compared to CON. Mundo Maya F1 hybrid was the sole high-yielding genotype apable of sustaining its yield under RR conditions. Our results suggested that its significant increase in fine root density (CON = 300 and RR = 910 root.m-2) and its maintenance of photosynthetic rate (2.5 - 3.5 mmol CO2 m-2 s-1) at high evaporative demand might explain why this genotype maintained high yield under RR condition.

Discussion: This work highlights a possible drought tolerance mechanism in fruit bearing adult coffee trees where the plant fine root number increases to intake more water in order to preserve turgor and sustainphotosynthesis at high ETo and therefore conserves high yield in dry conditions.

Keywords: Coffea arabica; Vietnam; climate change; drought tolerance; evapotranspiration; photosynthesis; plant physiology; water relations.

Figure 1

Experimental design of the trial studied in the northwest of Vietnam. (A) Flycam picture of part of the trial with the visible shade trees and rainfall reduction system (RR: upper part with vertical plastic bands and lower part: CON for control condition). (B) Detailed map of one repetition block of the trial with the 5 studied genotypes (CAT, Catimor; MAR, Marsellesa; STA, Starmaya; CEN, Centroamericano and MUN, Mundo Maya). The white arrow indicates field slope direction.

Figure 2

Environmental data across the experiment period. Rainfall and evapotranspiration (Rain & ETo) and soil water content in the first 50 cm (SWC-50). Dry season (D), wet season (W), transition period from dry to wet season (DW) and from wet to dry season (WD), in 2020 (20), 2021 (21) and 2022 (22). NA correspond to data not-available.

Figure 3

Performances of the C. arabica genotypes (Catimor [CAT], Marsellesa [MAR], Starmaya [STA], Centroamericano [CEN] and Mundo Maya [MUN]) tested under rainfall reduction (RR) or control (CON) conditions. The barplots presented the cumulated yield of 2020 & 2021 (A), root counts (B), longest plagiotropic branch length (C), and ΨPD (pre-dawn leaf water potentials) in April 2021 (D). Each dot is a data collected on one tree. Vertical bars in the bar plots are standard errors. Capital letters above isobars show significant differences between genotypes under CON, while lowercase letters show significant differences between genotype under RR treatment. Same letters mean no significant differences between genotypes. P-values show significant difference between the CON and RR treatments for a given genotype. Significance codes: 0 ***, 0.001 **, 0.01 *, 0.05 n.s, 1. n.s stands for non-significant.

Figure 4

(A) Photosynthesis (expressed in µmol CO₂ m^-2 LA s^-1), (B) conductance (expressed in mol H₂O m^-2 LA s^-1) in function of hourly VPD in kPa for the rain-fed (CON) vs rain reduction (RR) treatments of the 5 genotypes (CAT: Catimor, MAR: Marsellesa, STA: Starmaya, CEN: Centroamericano, MUN: Mundo Maya). Colored lines are simple linear models and grey areas are confidence intervals.

Figure 5

(A) Photosynthesis (µmol CO₂ m^-2 LA s^-1), (B) stomatal conductance (mol H₂O m^-2 LA s^-1) and (C) Whole-tree sapflow (mol H₂O m^-2 LA s^-1) of the 5 genotypes (CAT: Catimor, MAR: Marsellesa, STA: Starmaya, CEN: Centroamericano, MUN: Mundo Maya). Boxplots show a grand average of the photosynthesis (A1), conductance (B1) and whole-tree sapflow (C1) per genotype and time of day. Bold dots and the vertical line in the boxplot are the average and the standard error, respectively. Dots outside the boxplot are outliers. Lower case letters indicate Tukey-test significance with 95% confidence, genotypes with a same lower-case letter are not significantly different. Line graphs show the variation of photosynthesis (A2), conductance (B2) and whole-tree sapflow (C2) per genotype and time of day across seasons. Dry (D), wet seasons (W), transition periods (DW, WD) in 2020 (20), 2021 (21), and 2022 (22).

Figure 6

(A) Photosynthesis (expressed in µmol CO₂ m^-2 LA s^-1), (B) conductance (expressed in mol H₂O m^-2 LA s^-1) in function of hourly ETo in mm for the rain-fed (CON) vs rain reduction (RR) treatments of the 5 genotypes (CAT: Catimor, MAR: Marsellesa, STA: Starmaya, CEN: Centroamericano, MUN: Mundo Maya). Colored lines are simple linear models and grey areas are confidence intervals.