*3.1. Climate*

Rainfall accumulation during cover crop establishment was greater at all sites in the fall of 2016 compared to the fall 2017 (290 to 300 mm vs. less than 200 mm between September and November) (Figure 1). Both September and October were drier than average in southern France in 2017, while November and December were wetter. In April 2017 and May 2018, Arlington (site A), received more rain than average. The site received between 348.2 and 379.7 mm between April and June both in 2017 and in 2018 while the French sites only received between 173.4 and 216.4 mm over the same period. The greatest di fference in rainfall accumulation between Arlington and the French sites was observed in the summer. While Arlington (site A) received 198.6 to 278.6 mm between July and August of 2017 and 2018 the French sites only received 72.9 to 129.7 mm over the same period (Figure 1).

In Arlington in 2017, monthly average temperatures were below 0 ◦C from November to April. The coldest months were December and January, with a minimum air temperature mean of −16.3 ◦C. In 2018, the temperature raised above 0 ◦C a month later than in 2017 (early May vs. early April) and the coldest months were January and February with monthly minimum air temperatures of −12.8 and −11.5 ◦C, respectively. At the French sites, both winters were milder than in Wisconsin and periods of freezing temperatures were rare. In 2017, at site B, January was the coldest month with −1.5 ◦C on average. The monthly average temperature was above 10 ◦C from March to the end of the growing season. In 2018, February was colder than December and January with one week of frost. Meteorological stations close to the C, D and E sites indicated a monthly minimum air temperature of −1.2 to −0.7 ◦C in February 2018 compared with 4.5 to 5.5 ◦C in January. Monthly average temperatures were above 10 ◦C at the French sites at the beginning of April 2018.

**Figure 1.** Monthly rainfall accumulation and average temperature for each of the six field locations over the 2016–2017 and 2017–2018 seasons.

#### *3.2. Cover Crop Performance*

Data from the six trials analyzed with linear mixed models did not show any significant difference in biomass production between cereal rye and triticale, with 6989 kg ha−<sup>1</sup> and 7352 kg ha−1, respectively (Figure 2). However, cereal rye grew significantly taller than triticale, 125 cm vs. 77 cm, respectively (*p* < 0.001).

**Figure 2.** Mean weight of cereal rye and triticale biomass before cover crop rolling, averaged across all sites, 2017 and 2018. The linear mixed model did not indicate significant differences between cereal rye and triticale (*p* > *0.05*, *n* = 144). Data presented in Figure 2 are means ± standard error. Each cover crop species (rye and triticale) followed by the same letter are not significantly different. The dotted line refers to a mean of the cover crop biomass values range reported in the scientific literature as a success factor to suppress weed until soybean harvest.

Cover crop biomass of both cereal rye and triticale was highly influenced by the pedoclimatic conditions (location x year) and varied from 2,963 kg ha−<sup>1</sup> at the 18-Frce C site to 16,994 kg ha−<sup>1</sup> at 17-Arl A1. Except for 18-Arl A2, the ANOVA performed per site showed a significant effect of cover crop species on cover crop biomass. However, the cover crop species producing the highest biomass differed between sites: Cereal rye for 17-Frce B and 18-Frce D sites and triticale for 17-Arl A1, 18-Frce C and 18-Frce E sites (Figure 3). In 2017, at Arlington, triticale produced significantly greater biomass than any other the site, resulting in the nine outlying data points seen in Figure 2.

**Figure 3.** Cover crop biomass before termination per species and per site, 2017 and 2018. The letters represent the results of the ANOVAs per site (*p* < 0.05, *n* = 24). For each site, if the two species have the same letter their biomass is not significantly different. The dotted line refers to a mean of the cover crop biomass values range reported in the scientific literature as a success factor to suppress weed until soybean harvest.

#### *3.3. Weed Biomass*

In Wisconsin in 2018, the dominant weed species were Ladysthumb smartweed (*Polygonum persicaria* L.), lambsquarter (*Chenopodium album* L.), and foxtail (*Setaria pumila, Setaria viridis and Setaria faberi* L.). In France, the dominant weed species varied by location and year. At site B in 2017, common ragweed (*Ambrosia artemisiifolia* L.), heartsease (*Viola tricolor* L.) and Scarlet Pimpernel (*Anagallis arvensis*) dominated. In 2018, the weed population at site C was dominated by field bindweed (*Convolvulus arvensissuch* L.), all-seed (*Chenopodium polyspernum* L.) and round-leaved fluellin (*Kickxia spuria* L.) and at side D by yellow foxtail (*Setaria glauca* L.), switchgrass (*Panicum virgatum* L.) and persicaria (*Persicaria maculosa Gray* L.). Finally, in 2018 at site E, common ragweed was the main species along with annual bluegrass (*Poa annual* L.) and foxtail (*Setaria glauca* L.).

The linear mixed model determined that triticale provided poorer weed suppression compared to rye during the summer (Date 1), with 577 and 231 kg ha−<sup>1</sup> of weed biomass for triticale and cereal rye, respectively (Table 3). A similar conclusion was shown in the fall (Date 2), with 1545 and 1178 kg ha−<sup>1</sup> of weed biomass for triticale and rye, respectively. Weed dynamics between the summer and the fall did not differ significantly between rye and triticale. However, weed populations within the triticale cover crop tended to be higher than under cereal rye (*p* = 0.09) (Table 3). Indeed, data from the six trials indicated that the total weed biomass increased by an average of 945 kg ha−<sup>1</sup> for the rye and 968 kg ha−<sup>1</sup> for the triticale between Date 1 and Date 2.


**Table 3.** Summer and fall weed biomass from the two cover crop treatments over the six sites, 2017 and 2018. Weed biomass changes between the two dates, indicative of the degree of weed growth, are also reported.

1 Weed biomass was collected in July in France and in August at Arlington. 2 Weed biomass was collected in September. 3 Weed biomass mean from data of the six sites (*n* = 6 × 24) (17-Arl. A1, 18-Arl. A2, 17-Frce B, 18-Frce C, 18-Frce D, 18 Frce E) are presented in bold in the table for each cover crop specie Cereal Rye and Triticale at the different dates of measurement (Date 1 and Date 2). β Linear mixed model, n = 144, Significance codes: 0 '\*\*\*' 0.001 '\*\*' 0.01 '\*' 0.05 '.' 0.1 ' ' 1. Numbers in bold in the table followed by the different letters for Cereal Rye and Triticale within a similar date (Date 1, Date 2 or Date 2-Date 1) are significantly different.

The ANOVA per site showed that at the 18-Frce D site, the weed biomass was particularly high in the triticale plots with more than 4000 kg ha−<sup>1</sup> (Table 3). Significant differences between triticale and cereal rye were also observed during the summer (Date 1) at the 18-Arl. site with 1214 kg ha−<sup>1</sup> and 123 kg ha−<sup>1</sup> of weed biomass, respectively. 18-Frce C is the only site where cereal rye resulted in poorer weed suppression compared to triticale. However, at that site, increased weed biomass between Date 1 and Date 2 in the triticale was observed as compared to rye (Table 3).

In France, except for site B where weed development was limited, an increase in weed biomass of more than 1000 kg ha−<sup>1</sup> was observed between Date 1 and Date 2 for both cover crop species. At Arlington, weed biomass only increased by 204 kg ha−<sup>1</sup> (cereal rye) and 127 kg ha−<sup>1</sup> (triticale) between Date 1 and 2 in 2017 and decreased between the two measurements in 2018. Overall, the greatest increase in weed biomass between summer and fall was observed at 18-Frce C, D, and E (Table 3).
