European Journal of Agronomy 123 (2021) 126217 1161-0301/© 2020 Elsevier B.V. All rights reserved. Environment quality, sowing date, and genotype determine soybean yields in the Argentinean Gran Chaco Andr´ es Madias a, *, Guido Di Mauro b , Lucas N. Vitantonio-Mazzini b , Brenda L. Gambin b , Lucas Borr´ as b a AAPRESID - Asociaci´ on Argentina de Productores en Siembra Directa, Dorrego 1639 Piso 2 Ofcina A, S2000DIG, Rosario, Prov. de Santa Fe, Argentina b IICAR - CONICET, Concejo Nacional de Investigaciones Científcas y T´ ecnicas, Facultad de Ciencias Agrarias, Universidad Nacional de Rosario. Campo Experimental Villarino S/N, S2125ZAA, Zavalla, Prov. de Santa Fe, Argentina A R T I C L E INFO Keywords: Mixed-model effects Sowing date Years after land conversion Maturity group Yield limiting factors Deforestation ABSTRACT The Argentinean Gran Chaco is one of the global regions with the highest recent rates of agricultural expansion due to soybean production. The area has been heavily deforested during the last 30 years. Despite the economic relevance of soybean for this region, studies that provide options for sustainable management of these production systems are scarce. The objectives of this study were (i) to identify and to quantify key management and environmental soybean yield predictors, and (ii) to explore interactions with maturity group (MG) selection, since farmers are sowing genotypes ranging from MGs V to VIII. We evaluated commercial genotypes in 112 multi-environment on-farm trials (METs) consisting of 14–19 genotypes each during 11 consecutive years (from 2008 to 2019). We frst analyzed a single genotype sown in 106 METs to identify environmental and management yield predictors with good explanatory power for yield, which ranged from 435 to 5117 kg ha 1 . Relevant environ- mental variables were, in order of importance, rainfall from 30 days before sowing to physiological maturity (R7), years after land conversion to agriculture, reference evapotranspiration from sowing to R7, and the number of 2-day periods with maximum temperatures above 35 ◦ C from beginning of fowering (R1) to R7. Based on variable relative importance (RI) sowing date was the most important management variable (RI = 0.99), fol- lowed by phosphorous availability (RI = 0.59). Genotype selection also had a strong signifcant effect. There was an interaction between MG and sowing date, yield reductions with delayed sowings ranked as MG VIII > VII > VI = V. The largest yield differences between MGs were observed under environments with high soil organic matter (explored range from 1.64 to 4.05 %). These results illustrate specifc management variables to guide farmers and advisors to optimize regional soybean cropping systems. The negative yield effect promoted by years after land conversion and reductions in soil organic matter suggest a decline in the environmental quality of the region and the need for new production alternatives to halt these trends. 1. Introduction The global demand for soybean continues growing (Barret, 2019). This demand is met by increased production levels per cropped area but also by cropland expansion. Argentina currently produces 16 % of global soybean production (FAO, 2019), and it is the largest exporter of high protein soybean meal (USDA-FAS, 2020). This was possible because of both an increase in soybean acreage in the central temperate region (between latitudes 30 and 38 ◦ S) and the expansion of soybean into marginal areas (Aizen et al., 2009; Viglizzo et al., 2011; Ray et al., 2012). Northeastern of Argentina (integrated by Formosa, Chaco, eastern Santiago del Estero, and northern Santa Fe provinces) is part of the Gran Chaco region, and has experienced one of the highest rates of deforestation for agricultural purposes since the mid-90’s (Vallejos Abbreviations: AIC, Akaike’s information criterion; ω i , Akaike weights; BLUP, best linear unbiased predictor; MG, maturity group; ML, maximum likelihood; MMI, multimodel inference; METs, multi-environmental trials; PCV, proportional change in variance; R 2 C , conditional R 2 ; R 2 M , marginal R 2 ; RI, relative importance; REML, restricted maximum likelihood; VIF, variance infation factor. * Corresponding author. E-mail address: madias@aapresid.org.ar (A. Madias). Contents lists available at ScienceDirect European Journal of Agronomy journal homepage: www.elsevier.com/locate/eja https://doi.org/10.1016/j.eja.2020.126217 Received 12 April 2020; Received in revised form 19 November 2020; Accepted 24 November 2020