Dynamics of Spring Regrowth and Comparative Production Performance of 50 Autumn-Sown Alfalfa Cultivars in the Coastal Saline Soil of North China

Alfalfa (Medicago sativa L.) production is affected by many factors, including management practices, soil conditions, and the environmental elements of the target area. Varietal differences, in terms of agronomic performance and forage yield, among 50 alfalfa cultivars under six harvest systems following regrowth were evaluated during the growing season of 2019–2020 under non-irrigated rainfed conditions in a coastal saline-alkali soil region of North China. Days to harvesting, plant height, canopy area, growth rate, and forage yield were assessed to rank the cultivars. Furthermore, the key factor influencing the regrowth of the second year after over-wintering was identified based on the growth status before over-wintering by using the Boston Matrix method. Results showed significant (p < 0.05) differences among cultivars and harvests regarding plant height, canopy area, and forage yield. Alfalfa forage yield ranged between 24.2 t ha−1 yr−1 and 32.7 t ha−1 yr−1. The highest forage yield was obtained in cultivar Guochan No.1, and was lowest in cultivar Magnum 601. Forage yield reached the greatest values for the first harvest, and then decreased gradually and changed stably. The forage yield of the third, fourth, fifth, and sixth harvest ranged from 3.4 t ha−1 to 4.3 t ha−1 (averaged across 50 cultivars), which represented 10.8% to 15.2% of the annual total forage production. We also observed that forage yield correlated strongly, but negatively, with the growth rate. According to subordinate function value analysis, Womu No.1, WL440HQ, Weston, Surprise, and WL354HQ proved optimum cultivars for general cultivation in this coastal area. In future, development of alfalfa cultivars with improved regrowth and tolerance to heavy saline-alkali soil and early spring drought would be necessary to increase forage yield under rainfed conditions in coastal saline-alkali areas of North China.


Introduction
In recent years, livestock production has been increased rapidly, particularly in North China [1], which is highly dependent on adequate forage resources. According to the China Rural Statistical Yearbook [2], total livestock population in China is estimated to be 544.2 million pigs, 45.3 million cattle, 316.9 million sheep, and 14,640.6 million poultry. Grassland accounts for 40% of Chinese land, which has seen a loss of 22.7% from 1982 to 2010 [3]. The shortage of forage seriously restricts livestock husbandry. To ensure sufficient forage supplies, high quality forage of the high-yielding and salt and drought-tolerant cultivars is vigorously needed to be developed [4,5].

Description of Experimental Site
The field experiment was conducted in Nandagang County (117 • 22 E, 38 • 28 N) of Hebei Province in Northern China between August 2019 and October 2020. The study area belongs to a typical temperate continental monsoon-type climate. The average annual precipitation was 590 mm, which occurs mostly between June and September, while mean annual evaporation was 1950 mm. The groundwater table exists at approximately 0.3-1.2 m, with a salt concentration of 6 to 10 g L −1 . The soil in our experimental site had a silty clay loam texture. Weather data were collected from the China Meteorological Data Network, and the daily precipitation and maximum and minimum temperature during 2019 and 2020 are presented in Figure 1. In 2019, total precipitation was 556.6 mm and maximum and minimum mean air temperature was 9.5 • C and 19.7 • C, respectively. In 2020, total precipitation was 748.0 mm and maximum and minimum mean air temperature was 9.6 • C and 19.3 • C, respectively.
Life 2021, 11, x FOR PEER REVIEW 3 of 22 yield under rainfed conditions, in order to facilitate the selection of high-yielding cultivars that adapted to this study area.

Description of Experimental Site
The field experiment was conducted in Nandagang County (117°22′ E, 38°28′ N) of Hebei Province in Northern China between August 2019 and October 2020. The study area belongs to a typical temperate continental monsoon-type climate. The average annual precipitation was 590 mm, which occurs mostly between June and September, while mean annual evaporation was 1950 mm. The groundwater table exists at approximately 0.3-1.2 m, with a salt concentration of 6 to 10 g L −1 . The soil in our experimental site had a silty clay loam texture. Weather data were collected from the China Meteorological Data Network, and the daily precipitation and maximum and minimum temperature during 2019 and 2020 are presented in Figure 1. In 2019, total precipitation was 556.6 mm and maximum and minimum mean air temperature was 9.5 °C and 19.7 °C, respectively. In 2020, total precipitation was 748.0 mm and maximum and minimum mean air temperature was 9.6 °C and 19.3 °C, respectively.

Experimental Design and Field Management
A total of 50 alfalfa cultivars were tested for agronomic performance and forage yield using randomised block design (RBD) with three replications. These 50 cultivars represented themselves as 50 different treatments, and the details regarding the dormancy level, origin, and seed source of each cultivar is presented in Table 1. Each alfalfa cultivar

Experimental Design and Field Management
A total of 50 alfalfa cultivars were tested for agronomic performance and forage yield using randomised block design (RBD) with three replications. These 50 cultivars represented themselves as 50 different treatments, and the details regarding the dormancy level, origin, and seed source of each cultivar is presented in Table 1. Each alfalfa cultivar was planted in a separate plot with a net size of 16 m 2 (4 m × 4 m). Sowing was done on 20 August 2019. A basal dose of 70.1 kg N ha −1 as urea and 80.7 kg P 2 O 5 ha −1 as diammonium phosphate was applied just before the seed sowing on the day of planting. Afterwards, 22.7 kg ha −1 potassium dihydrogen phosphate was continuously applied after was harvested when alfalfa reached 20% bloom stage. At each harvest, aboveground plants were mowed at the height of 5 cm from the ground level to facilitate fast regrowth.

Sampling and Measurements
To monitor the plant height of alfalfa (AH, cm), five plants from five different points in each plot were randomly measured from the ground level to the top of the plants using a ruler every 5 days. Growth rate was calculated using the formula (AH 2 − AH 1 )/(d 2 − d 1 ), where AH 2 and AH 1 represent the plant height measured at the day d 2 and d 1 , respectively. Canopy projection area (CPA) affects alfalfa growth dynamics. The CPA was measured when alfalfa was growing vigorously (CPA = length × height). Three 0.7 × 1 m quadrat were taken from each experimental plot to estimate fresh biomass yield. The aboveground dry biomass of harvested samples from each plot was determined by oven-drying at 70 • C until a constant weight was reached. Forage yield of alfalfa was calculated on dry-weight basis. Total annual yield for each cultivar was determined by adding the biomass yield from all six harvests.

Boston Matrix Method
The difference of alfalfa growth was analysed by using two parameters, namely growth status (forage yield, plant height, canopy area, and growth rate). These two parameters were used to reflect the effects of growth status before over-wintering (when alfalfa was planted in autumn in different cultivar) on regrowth in the second year by applying the Boston Matrix method [28].

Subordinate Function Value
Production adaptability involving multiple indices i.e., plant height, canopy area, growth rate, and forage yield, were evaluated based on their subordinate function value (SFV) [29] which is shown in formula (1) as following: where X is the comprehensive value of each index of alfalfa cultivars; X max and X min are maximum and minimum values for comprehensive index, respectively. All the data were presented as mean ± standard deviation (SD) of three replications. Differences among treatments were separated by the least significance difference (LSD) test at 0.05 significance level using the SAS (8.0) software package. Correlations and regression analysis were performed using SPSS 17.0. The figures were plotted using SigmaPlot 12.5. Boston Matrix analysis was performed using MS Excel 2020.

Alfalfa Growth Period and Forage Yield
Differences were observed in the alfalfa growth period of all 50 cultivars ( Table 2). The seedling emergence time ranged between 1 and 6 days, with the first germinations appearing in cultivars AC4, AC14, AC19, and AC27. The earliest flowering appeared in the cultivars AC3, AC12, AC13, AC14, AC21, AC22, AC31, and AC32, and the variation ranged from 1 to 5 days. The vegetative stage was the longest in cultivar AC10, while shortest vegetative stage was observed in cultivars AC20, AC21, AC22, AC31, AC32, AC34, AC39, and AC40.
Overall, the forage yield of alfalfa from the first to sixth cutting decreased significantly (p < 0.05), and tended to adapt to a gradual decreasing trend (Figures 2 and 3). The highest forage yield was recorded in the first cutting (7.2 t ha −1 ), while the lowest was recorded in the third cutting (3.1 t ha −1 ), accounting for 25.2% and 10.8%, respectively, of the total annual forage yield. Results further indicated that the effect of alfalfa cultivars on total annual forage yield was highly significant (p < 0.05; Table 3). Cultivars AC1, AC13, AC7, AC5, and AC21 ranked in the top five regarding the total annual forage yield, which produced forage yields of 32.7 t ha −1 , 31. respectively. The correlation analysis between the total annual yield and agronomic traits showed that the total annual yield was not associated with plant height and growth rate, but it was negatively correlated with the growth rate (r = −0.300, p < 0.05) ( Table 4). On the other hand, plant height was positively associated with both canopy area and growth rate (r = 0.926, p < 0.01; r = 0.312, p < 0.05; respectively). AC7, AC5, and AC21 ranked in the top five regarding the total annual forage yield, which produced forage yields of 32.7 t ha −1 , 31.7 t ha −1 , 31.4 t ha −1 , 31.4 t ha −1 , and 31.3 t ha −1 , respectively. The correlation analysis between the total annual yield and agronomic traits showed that the total annual yield was not associated with plant height and growth rate, but it was negatively correlated with the growth rate (r = −0.300, p < 0.05) ( Table 4). On the other hand, plant height was positively associated with both canopy area and growth rate (r = 0.926, p < 0.01; r = 0.312, p < 0.05; respectively).       As shown in Figure 4, there were 15 cultivars in Category I, namely, WL440, Ameri graze 37 CR, and Magnum551, etc. The forage yield before over-wintering was not high, but their total forage yield in the second year was higher. There were 15 cultivars located in Category II, including Magnum601, SK4020, Phabulous, and Magnum7, etc. These cultivars were not good both in forage yield before over-wintering and in the second year. Similarly, there were 15 cultivars in Category III, namely Seniti, ladino, WL903HQ, and WL298HQ, etc. These cultivars had a higher forage yield before over-wintering and a higher total forage yield in the second year. Furthermore, there were five cultivars in Category IV, namely Magnum7, Eureka, WL656HQ, Khan, and SK3010. They had a higher forage yield before over-wintering, but low total forage yield during the second year.   As shown in Figure 4, there were 15 cultivars in Category I, namely, WL440, Ameri graze 37 CR, and Magnum551, etc. The forage yield before over-wintering was not high, but their total forage yield in the second year was higher. There were 15 cultivars located in Category II, including Magnum601, SK4020, Phabulous, and Magnum7, etc. These cultivars were not good both in forage yield before over-wintering and in the second year. Similarly, there were 15 cultivars in Category III, namely Seniti, ladino, WL903HQ, and WL298HQ, etc. These cultivars had a higher forage yield before over-wintering and a higher total forage yield in the second year. Furthermore, there were five cultivars in Category IV, namely Magnum7, Eureka, WL656HQ, Khan, and SK3010. They had a higher forage yield before over-wintering, but low total forage yield during the second year.

Plant Height
There were significant differences in plant heights during different cutting dates (p < 0.05), and the order was as following: first cutting > sixth cutting > fifth cutting > fourth

Plant Height
There were significant differences in plant heights during different cutting dates (p < 0.05), and the order was as following: first cutting > sixth cutting > fifth cutting > fourth cutting > third cutting > second cutting ( Figure 5). When compared with the first cutting, the mean plant height of alfalfa in the second, third, fourth, fifth, and sixth cutting decreased by 47.4%, 31.4%, 29.5%, 28.2%, and 24.4%, respectively. The overall mean plant height for all 6 cuttings ranged from 29.9 to 62.0 cm. The tallest cultivars were AC25 (62.0 cm) and AC14 (61.7 cm), whereas the shortest cultivar was AC10 with a height of 29.9 cm (p < 0.05) ( Table 5).
As shown in Figure 6, there were 19 cultivars in Category I, namely, WL712HQ, WL440HQ, Eureka, and Vision, etc. Their plant heights before over-wintering were not high, but their mean plant heights in the second year were higher. There were 20 cultivars located in Category II, including WL903HQ, WL298HQ, Magnum7, and 4015, etc. These cultivars were not good, both in plant height before over-wintering and during the second year after over-wintering. There were seven cultivars in Category III, namely Magnum401, Alfalfa, Low TA, and Magnum7, etc. These cultivars had a higher plant height and a higher mean plant height in the second year. Furthermore, there were four cultivars in Category IV, namely Australia Queen, WL168HQ, SK4020, and Khan. They had higher plant heights before over-wintering, but low mean plant heights during the second year.

Canopy Area
Canopy area had significant variation among the six harvests when averaged across all cultivars ( Figure 5). The canopy area of the first cutting was significantly higher (p < 0.05) than rest of the cuttings, followed by the fifth cutting. The order of canopy area was: first cutting > fifth cutting > fourth cutting > third cutting > sixth cutting > second cutting. Compared with the first cutting, canopy area of alfalfa in second, third, fourth, fifth, and sixth cuttings decreased by 63.8%, 36.9%, 34.8%, 32.2%, and 52.7%, respectively. Mean canopy area also varied significantly (p < 0.05) among cultivars, with the greatest mean canopy area measured for cultivars AC13, AC14, AC16, AC21, AC25, and AC40 (0.24 m 2 ), while the smallest canopy area was measured for cultivar AC10 (0.09 m 2 ) ( Table 5).

Canopy Area
Canopy area had significant variation among the six harvests when averaged across all cultivars ( Figure 5). The canopy area of the first cutting was significantly higher (p < 0.05) than rest of the cuttings, followed by the fifth cutting. The order of canopy area was: first cutting > fifth cutting > fourth cutting > third cutting > sixth cutting > second cutting. Compared with the first cutting, canopy area of alfalfa in second, third, fourth, fifth, and sixth cuttings decreased by 63.8%, 36.9%, 34.8%, 32.2%, and 52.7%, respectively. Mean canopy area also varied significantly (p < 0.05) among cultivars, with the greatest mean canopy area measured for cultivars AC13, AC14, AC16, AC21, AC25, and AC40 (0.24 m 2 ), while the smallest canopy area was measured for cultivar AC10 (0.09 m 2 ) ( Table 5).
As shown in Figure 7, there were 23 cultivars in Category I, namely, Athena, Ladino, WL903HQ, and WL440, etc. The canopy area before over-wintering was not high, but their mean canopy area in the second year after over-wintering was higher. There were 10 cultivars located in Category II, including Guochan No. l, Magnum401, Australia Queen, and Low TA, etc. These cultivars were not good, both in canopy area before over-wintering and during the second year. There were 14 cultivars in Category III, namely WL366HQ, WL712HQ, Magnum7, and AC caribou, etc. These cultivars had a higher canopy area and a higher mean canopy area in the second year. Likewise, there are three cultivars in Category IV, namely WL168HQ, SK4030, and Khan. They had higher canopy area before overwintering, but a low-mean canopy area during the second year. As shown in Figure 7, there were 23 cultivars in Category I, namely, Athena, Ladino, WL903HQ, and WL440, etc. The canopy area before over-wintering was not high, but their mean canopy area in the second year after over-wintering was higher. There were 10 cultivars located in Category II, including Guochan No. l, Magnum401, Australia Queen, and Low TA, etc. These cultivars were not good, both in canopy area before over-wintering and during the second year. There were 14 cultivars in Category III, namely WL366HQ, WL712HQ, Magnum7, and AC caribou, etc. These cultivars had a higher canopy area and a higher mean canopy area in the second year. Likewise, there are three cultivars in Category IV, namely WL168HQ, SK4030, and Khan. They had higher canopy area before over-wintering, but a low-mean canopy area during the second year.

Growth Rate
In the second cutting, the growth rate was significantly higher than the other cutting dates, and the sequence was as following: second cutting > fourth cutting > third cutting > fifth cutting > first cutting > sixth cutting ( Figure 5). The growth rate of alfalfa in the first, third, fourth, fifth, and sixth decreased by 49.8%, 24.6%, 15.5%, 51.1%, and 50.8%, respec-

Growth Rate
In the second cutting, the growth rate was significantly higher than the other cutting dates, and the sequence was as following: second cutting > fourth cutting > third cutting > fifth cutting > first cutting > sixth cutting ( Figure 5). The growth rate of alfalfa in the first, third, fourth, fifth, and sixth decreased by 49.8%, 24.6%, 15.5%, 51.1%, and 50.8%, respectively, when compared with the second cutting. No significant differences in growth rates were observed among alfalfa cultivars during all cuttings ( Table 5). The mean growth rate of alfalfa reached the maximum of 2.5 cm d −1 in AC43, AC44, and AC50 and the minimum of 1.3 cm d −1 in AC10.
As shown in Figure 8, there were 15 cultivars in Category I, namely, Magnum7, Algonguin, Surprise, and WL363HQ, etc. The growth rate before over-wintering was not high, but their mean growth rate in the second year was higher. There were 17 cultivars located in Category II, including Inster, Alfalfa, Tango, and Low TA, etc. These cultivars were not good, both in growth rate before over-wintering and during the second year. There are nine cultivars in Category III, namely WL525HQ, Australia Queen, Seniti, and Weston, etc. These cultivars had a higher growth rate and a higher mean growth rate in the second year. Lastly, there were five cultivars in Category IV, namely FDA No.1, WL168HQ, WL366HQ, and WL903HQ. They had higher growth rates before over-wintering, but low-growth rates during the second year.

Comprehensive Evaluation
The subordinate function value (SFV) of all cultivars varied from 1.50 to 2.89; the maximum and minimum SFVs were observed in cultivar AC11 and AC8, respectively (Table 6). Among 50 alfalfa cultivars, five cultivars, AC8, AC25, AC26, AC48, and AC42, had better production adaptability. Meanwhile, cultivars AC11, AC10, AC43, AC4, AC47, and AC15 showed the lowest production adaptability. The distribution of SFV reflected the mean value and standard deviation of 2.04 and 0.26, respectively (Figures 9 and 10).   Cluster analysis based on mean plant height, canopy area, growth rate, and total annual yield classified the tested 50 alfalfa cultivars into three groups ( Figure 11). Cluster I was the group with highest plant height and canopy area, consisting of cultivars AC29, AC30, AC50, and AC42. The top five cultivars in terms of total annual yield were AC1, AC13, AC7, AC5, and AC21, and they were also ranked in Cluster I. Cluster II was the group with highest growth rate, consisting of cultivars AC11, AC24, AC27, AC12, AC41,  Figure 9. Percentage proportion of each harvest to total annual yield of different alfalfa cultivars. Figure 9. Percentage proportion of each harvest to total annual yield of different alfalfa cultivars. Cluster analysis based on mean plant height, canopy area, growth rate, and total annual yield classified the tested 50 alfalfa cultivars into three groups ( Figure 11). Cluster I was the group with highest plant height and canopy area, consisting of cultivars AC29, AC30, AC50, and AC42. The top five cultivars in terms of total annual yield were AC1, AC13, AC7, AC5, and AC21, and they were also ranked in Cluster I. Cluster II was the group with highest growth rate, consisting of cultivars AC11, AC24, AC27, AC12, AC41, and AC15. The cultivar AC10 had the lowest canopy area and growth rate, and thus it was placed into Cluster III.

Forage Yield and Yield Components
Plant height, canopy area, growth rate, and forage yield in this study indicated that all tested cultivars had active growth during winter in this coastal region (Figures 2 and 3). Alfalfa is usually harvested two to four times each year in the northern regions of China [30]. However, we found that a sixth harvest could be achieved in the coastal saline-alkali soil of North China and the highest forage yield was found in the first harvest. The reported annual alfalfa yield in our study (28.5 t ha −1 ) is higher than that reported in other studies conducted in the same region. Alfalfa annual forage yield varied from 10.6 to 14.4 t ha −1 in the Hebei area [31], and a low yield of 12 alfalfa cultivars with a maximum annual forage yield of 20.1 t ha −1 observed in Cangzhou City [32]. The forage yield of the Xunlu, WL656HQ, WL712HQ, WL363HQ, WL525HQ, and WL903HQ were 12.4, 12.6, 13.3, 12.4, 12.9, and 13.5 t ha −1 in study region, respectively [31]. These differences in forage yield may be due to variations in harvest frequencies and environmental conditions. The mean plant height and canopy area of all cultivars at the first cutting were significantly taller (78.2 cm and 0.33 m 2 , respectively) than at all other cuttings. This could be due to the reason that cultivars took enough time (132 days) to achieve maximum growth before the second cutting, while all other cuttings got a shorter time (3-30 days) for growth and development compared with the first cutting. In the present study, canopy area and growth rate were strongly associated with forage yield, which implies that the level of canopy cover could represent the aerial biomass [33]. The above results indicate that the forage yield of alfalfa is positively correlated to canopy cover, and this could reduce temperature and transpiration [33][34][35]. Cluster Ⅰ Cluster Ⅱ Cluster Ⅲ Figure 11. Cluster analysis for production adaptability of 50 alfalfa cultivars based on plant height, canopy area, growth rate, and mean annual yield.

Major Factors Influencing Production Adaptability of Alfalfa
Many factors affect the production adaptability of alfalfa including plant genotype, environmental conditions, field management practices, and stage of maturity [36,37]. The yield reduced by 42% in rainfed cropping compared to irrigation cropping [38]. Similar results were reported in another study [39]. Additionally, favourable air temperature and high precipitation during the alfalfa growing period contributed to the high alfalfa yield [12,40]. Drought and salt-stress-associated defoliation and diseases during summer resulted in plant death [41]. Air temperature and rainfall appeared to be major determinants on the growth of alfalfa cultivars in this study (Table 2), and the timing of consecutive harvests was dependent on the weather conditions during the alfalfa growing period [18]. This is not surprising because the major part of alfalfa growth in this study region occurs from May to September. This also emphasised the importance of rainfall on forage yield during the alfalfa growing period [21,42]. High temperature reduced forage yield, potentially because of reduced soil moisture due to high evaporation [21]. In this study, the precipitation amounts and mean temperatures were: 28.2, 5.6, 158.5, 324.5, and 73.7 mm, and 23.0, 26.9, 26.9, 27.0, and 22.9 • C from harvest one to six, respectively ( Figure 1). The forage yield in the third cutting was the lowest, which might be related to the low precipitation. As coastal regions are usually characterised by high salinity and irregular rainfall, drought-and salt-tolerant cultivars should be selected for planting in order to ensure optimum forage yield. Alternating temperature, rainfall, and cultural practices may have a great impact on alfalfa growth, which ultimately results in the variation of the alfalfa yield. Further adoption of cultural practices aimed to improve forage yield may therefore be desired, including new cultivation models of inter-planting alfalfa with different crops for the decreasing of alfalfa canopy temperatures, and the further development and release of drought-tolerant cultivars for this study region.
Alfalfa forage yield is significantly influenced by the harvest intervals or cutting frequencies. Moreover, alfalfa growth is sensitive to the amount of growth days, and a short growing period would ultimately result in lower dry matter accumulation, which can be observed in the 23-30 days inter-harvest intervals for the second, third, fourth, fifth, and sixth cuttings in this study (Table 2). Forage yields with between 30-and 40-day harvest intervals were higher than the forage yield harvested with a 20-day interval [23]. Similar results regarding forage yield as being affected by harvest interval time were also reported [24,43]. In our study, although growth rate was the highest in the second harvest, with 23 days to reach blossom, forage yield was smaller than that in the first harvest. Similarly, the growth period was shorter from the third to sixth harvests than in other harvests which resulted in a smaller forage yield. Other researchers also noted a positive relationship between the forage yield and the regrowth period [44]. Therefore, in the coastal saline soil of Northern China, researchers should focus on the development of effective management practices for increasing alfalfa growth from the second to sixth harvests, which is of great importance to increasing the forage yield under such climate conditions.
Additionally, regression analysis of the predicted fall dormancy rating vs. plant height and canopy area in this study resulted in high R 2 values and positive slopes (plant height, y = 0.8922x + 51.42, R 2 = 0.126 **; canopy area, y = 0.0063x + 0.1723, R 2 = 0.2288 **) ( Figure 12). However, the fall dormancy rating did not influence forage yield and growth rate, indicating that no effects of fall dormancy levels were observed on the forage yield in this study. In the present study, the three alfalfa cultivars with the highest annual yields ranging from 30.1 to 31.3 t ha −1 among fall dormancy ratings of 9-10 were selected from the evaluation of 50 cultivars. The main reason for this was the relatively high temperature in study region (Figure 1). The similar results were also obtained from another study [45]. The results of the fall dormancy ratings of the alfalfa cultivars were consistent with several previous studies [22,46,47]. rate, indicating that no effects of fall dormancy levels were observed on the forage yield in this study. In the present study, the three alfalfa cultivars with the highest annual yields ranging from 30.1 to 31.3 t ha −1 among fall dormancy ratings of 9-10 were selected from the evaluation of 50 cultivars. The main reason for this was the relatively high temperature in study region (Figure 1). The similar results were also obtained from another study [45]. The results of the fall dormancy ratings of the alfalfa cultivars were consistent with several previous studies [22,46,47]. Figure 12. Regression analysis of fall dormancy rating vs. plant height, canopy area, growth rate, and mean annual yield during the six harvests. Coefficient of determination (R 2 ) designated with "**" is significantly different (p < 0.05).

Conclusions
In the present study, 50 alfalfa cultivars achieved very satisfactory total forage yields (ranging from 24.2 to 32.7 t ha −1 ) through their planting under non-irrigated rainfed conditions (748 mm precipitation and 14.4 °C temperature in the study area) in a coastal saline-alkali soil region. Six harvests of alfalfa could be achieved in this study region; however, the highest forage yield was found in the first harvest. We also found that there was great potential to improve the alfalfa forage yield from the second to sixth harvests because of the influence of air temperature and precipitation during the alfalfa growth period from May to September. Moreover, forage yield was negatively correlated with the growth rate. Importantly, cultivar Womu1 showed superior agronomic performance in Figure 12. Regression analysis of fall dormancy rating vs. plant height, canopy area, growth rate, and mean annual yield during the six harvests. Coefficient of determination (R 2 ) designated with "**" is significantly different (p < 0.05).

Conclusions
In the present study, 50 alfalfa cultivars achieved very satisfactory total forage yields (ranging from 24.2 to 32.7 t ha −1 ) through their planting under non-irrigated rainfed conditions (748 mm precipitation and 14.4 • C temperature in the study area) in a coastal saline-alkali soil region. Six harvests of alfalfa could be achieved in this study region; however, the highest forage yield was found in the first harvest. We also found that there was great potential to improve the alfalfa forage yield from the second to sixth harvests because of the influence of air temperature and precipitation during the alfalfa growth period from May to September. Moreover, forage yield was negatively correlated with the growth rate. Importantly, cultivar Womu1 showed superior agronomic performance in terms of growth and forage yield. Future studies need to be conducted which focus on improving forage yields by adopting the rational cultivation model and inter-planting with different crops under optimal temperatures and water conditions in the saline soil regions.

Data Availability Statement:
The authors confirm that the data supporting the findings of this study are available within the article.