Weather conditions

The climatic conditions of the production cycles and of the last 30 years (1985-2014) are shown in Figure 1. The minimum, average and maximum temperatures of the 2012-2013 production cycle were higher in November, December and January compared to the 2013-2014 production cycle and to the average of the 30 years, although the average maximum temperature of the 30 years in January was slightly higher than that of the two production cycles in this month. The minimum, average and maximum temperatures of February, March and April were higher in the 2012-2013 production cycle than the temperatures of these months on average of the 30 years, but with slight increases in the minimum and maximum temperatures in March during the average 30 years.

Figure 1 Monthly temperature and rainfall during the growing season of forage crops in two production cycles (2012-2013 and 2013-2014) and the average 30 years (1985-2014) in the La Laguna Experimental Field. 

However, higher minimum, average and maximum temperatures occurred during February, March and April in the 2013-2014 production cycle than in the 2012-2013 production cycle and the average 30 years; except for an increase in the maximum temperature in March during the average 30 years when compared to the two production cycles. Regarding rainfall, there was a higher total seasonal rainfall (November to April) in the 2013-2014 production cycle (59.6 mm) than in the 2012-2013 production cycle (6.6 mm) and the average 30 years (38.2 mm). Overall, the climate during the growing season of the crops was warmer and drier in the 2012-2013 production cycle than during the 2013-2014 production cycle and the average 30 years.

Nutrient composition and nutrient yield in the 2012-2013 and 2013-2014 production cycles

The interaction species × sowing date affected DM yields. During the 2012-2013 production cycle, the barley sown on December 7 and the triticale sown on November 16 and December 28 had the highest DM yields. However, the DM yield of the triticale sown on December 28 was similar to those observed in the barley sown on November 16 and December 28 and that of the triticale sown on December 7 (Table 1). During the 2013-2014 production cycle, the barley sown on November 16 had a higher DM yield than the other species on any sowing date.

In addition, the barley sown on December 7 had a DM yield similar to that of the triticale sown on December 7 and 28. The lowest DM yield occurred for canola and safflower sown on December 28 and January 18. The interaction species × sowing date also indicated that the highest yields of NE_L occurred in barley and triticale sown on November 16 and December 7 in both production cycles (p< 0.0001; Tables 1 and 2). The triticale sown on November 16 had yields of NE_L similar to those obtained by the barley sown on December 7, but in the second year, the barley sown on November 16 was superior to the triticale sown on any sowing date. This behavior could indicate that triticale yields better in warmer and drier years than barley (^{Roohi et al., 2013}).

DM and nutrient yields decreased as the sowing date was delayed in all species, although in triticale this effect was not as pronounced. This was probably because this species obtained the highest yields on December 28 and yields similar to or higher than those of barley on January 18. Several studies have found a reduction in DM yield as the sowing date is delayed, either in barley (^{Mirosavljević et al., 2018}), triticale (^{Schwarte et al., 2005}), canola (^{Sánchez et al., 2018}) and safflower (^{Talji, 2015}).

However, these effects are commonly associated with the reduction of the number of productive tillers or branches in barley and safflower (^{Mirosavljević et al., 2018}), as well as shortening the growing season in canola (^{Begna and Angadi, 2016}) and triticale (^{Schwarte et al., 2005}), which restricts leaf area and the interception of solar radiation by delaying the sowing date. Plant growth and development are directly or indirectly influenced by environmental variables such as photoperiod, temperature and water content in the soil. In the present study, the photoperiod was the same in the two growth cycles since the same sowing dates were evaluated; therefore, differences in DM yield between years were mainly related to temperature and precipitation during the growth cycle.

The effects of heat stress on plants are very complex, causing alterations in growth and development with changes in physiological functions such as reduced photosynthetic capacity and decreased metabolic activity, which is reflected in the reduction of grain formation and yield (^{Nurunnaher and Islam, 2017}). In the present study, the DM yield was higher in the 2013-2014 cycle (8 547 kg ha^-1; Table 2) compared to that observed in 2012-2013 (8 147 kg ha^-1; Table 1).

This behavior was associated with the lower maximum and minimum temperature that occurred in 2013-2014, mainly in November, December and January (Figure 1), which led to a longer growth cycle (5 days) compared to the 2012-2013 cycle. In addition, in 2013-2014, there was greater rainfall, which occurred in November, December and January (Figure 1), which was a complement to the irrigation applied to crops, mainly during the vegetative period of the first two sowing dates. In this study, the differences in DM yield between sowing dates are also related to variations in photoperiod and environmental temperature during the period of crop development.

The difference in photoperiod between sowing dates was up to 0.5 hours, while in the average maximum temperature during the cycle, the increase was 3.9°C between the sowing date of November 16 and January 18 (Figure 1). Both factors influenced the length of the growth cycle, with an acceleration of the cycle in all crops of 17 to 29 days when sowing was delayed from November 16 to January 18. The decrease in DM yield as the sowing date was delayed was the result of a lower absorption of solar radiation and its resulting lower accumulation of DM. In addition, the prevailing temperature level during the cycle affected the growth of the crops.

The optimal temperatures for the growth of canola, safflower and barley are 20 to 23°C, and 25.6°C for triticale; while the maximum temperatures are 30°C for canola, 37°C for barley and 40°C for triticale and safflower (^{Robertson et al., 2002}; ^{Gol et al., 2017}; ^{Munyon et al., 2020}; ^{Torabi et al., 2020}). The highest DM yields on the early sowing dates of November 16 and December 7 coincide with the average maximum temperatures (24.8 to 25.05°C) closest to optimum temperatures; whereas, in the late dates of December 28 and January 18, temperatures increased to 26.3 and 27.9°C, respectively (Figure 1), which are closer to maximum temperatures, especially in canola.

CP and NDF yields were also affected by the interaction species × sowing date in both production cycles. During 2012-2013, canola sown on November 16 and December 7 showed the highest CP yield. These yields were equal to those obtained in barley on December 7, safflower on November 16 and December 7 and triticale sown on November 16 and December 28. The lowest CP yields occurred in triticale sown on December 7 and January 18, safflower on January 18 and barley and canola in sowings of December 28 and January 18 (p< 0.0001).

Regarding NDF yield, the highest values were observed in the barley sown on December 7 and the triticale on November 16. These yields were equaled by the yields of the barley sown on November 16 and December 28 and the triticale on December 7 and 28. The lowest NDF yield was observed in canola sown on November 16, December 28 and January 18 and that of safflower sown on December 28 and January 18 (p= 0.003). During 2013-2014, the highest CP yield was obtained in canola and safflower sown on November 16, and this was similar to that obtained with barley on November 16 and safflower on December 7.

NDF yield was higher in barley and triticale sown on November 16 and December 7, respectively. This was similar to that found in barley on December 7 and triticale on December 28. The lowest NDF yield was observed in canola and safflower sown in December and January (p< 0.0001). Due to the minimal variation in nutrient composition by sowing date, this suggests that the best yields of these nutrients were observed when sowing between November 16 and December 7. A similar response was observed in a study in which the nutrient production of six canola cultivars sown on four sowing dates was determined (^{Sánchez et al., 2018}). The results indicated that the CP yields produced on the sowing date of October 19 were higher than those produced on September 2 and 19.

As in our study, these results were attributed to the highest DM yields obtained on the late sowing date, when lower environmental temperatures and short photoperiod prolonged the canola growth cycle. The percentages of CP, ADF and NDF of the forages were affected by the species and sowing date in the two production cycles. In 2012-2013, canola was superior to all species in CP, followed by safflower, barley and triticale. For the sowing date, the highest CP was observed in sowings of January 18. Triticale and barley had the highest ADF concentrations and the highest ADF was observed on the sowing dates of December 7 and 28 (Table 1).

In 2013-2014, canola and safflower had the highest concentration of CP and the optimal sowing date to obtain the highest concentration of CP in canola, safflower and barley was January 18. Regarding fiber content, barley and triticale had a higher concentration of NDF than canola and safflower. The highest concentration of NDF was observed when forages were sown from December 7 to January 18 (Table 2).

The percentage of ADF and the concentration of NE_L in the forages were affected by the species during 2013-2014 (Table 2). Barley and triticale had a higher concentration of ADF than canola and safflower. However, barley and triticale had lower NE_L concentrations than canola and safflower (p< 0.0001). The results indicate that the greatest significant effect was due to the species; although the nutrient composition of the forage showed variations by the date of sowing, by the species or by the interaction between the two.

CP content was on average 7 percentage units higher in canola and safflower than in barley and triticale in both production cycles. Results similar to those of the present study were found when comparing alternative autumn-winter forages with traditional forages (^{Reta et al., 2017b}). In addition, a higher concentration of CP and a lower yield of DM obtained in all forage species grown on the late sowing date suggest that the plants acquired excess nitrogen, while the accumulation of DM was probably limited to the warmer days during plant development. The latter has been described in studies with ryegrass species grown in summer (^{Santiago et al., 2012}).

As for the percentage of fiber, NDF was lower in canola and safflower compared to barley and triticale. ^{Reta et al. (2017b}) reported 11% more NDF in barley and triticale forages than in canola and safflower. In the present study, the concentration of NE_L in canola and safflower obtained in the second year was higher than that observed for barley and triticale; however, this result was not observed in the first year, which was probably associated with higher environmental temperatures. This plant response is usually associated with increased cell wall thickness in response to high temperatures (^{Li, 2021}).

In our study, warmer temperatures increased NDF and ADF contents while reducing the concentration of NE_L in canola and safflower forage. Similar results have been found in canola, in which high environmental temperatures reduced CP concentration and increased NDF and ADF contents (^{Wiedenhoeft and Barton 1994}). ^{Reta et al. (2016)} also indicated lower NE_L contents and higher concentrations of NDF and ADF in canola grown under high environmental temperatures.

In this study, the best sowing date in canola and safflower, considering its nutrient composition and nutrient yield, has been found. However, it is necessary to carry out further research that considers the digestibility of the fiber and its production per hectare, as well as the best conservation method for these alternative forages.