Variation of Fertility and Phenological Synchronization in Cunninghamia lanceolata Seed Orchard: Implications for Seed Production

Abstract

Reproductive synchronicity between parents influences the seed production and quality in seed orchards. Our objective was to determine clonal variation in fertility and phenological synchronization, as well as their effect on seed production, in a Chinese fir (Cunninghamia lanceolata) open-pollinated seed orchard. Significant variation of female fertility and male phenological synchronization occurred in the clones. The flowering of the male was 2 days earlier than the female. The gamete contribution of female and male were unbalanced between clones (the phenological synchronization indexes (PO_ij) were 0.000–0.585 (as female) vs. 0.000–0.385 (as male)). In general, the average PO_ij value of as a male was lower than as a female, but the number of male flowers were significantly higher than female, indicating that the management of female flowers should be enhanced. The average PO_ij of self-pollination was 0.298, higher than cross-pollination (0.236), indicating that there was a larger probability to selfing in the orchard. The middle phenological type possessed higher phenological synchronization indexes than early and late phenological type. Genetic control was stronger for number of female flowers (H² = 0.277) than for male and female initial stages and flower duration (H² = 0.193–0.239). We found a positive correlation between PO_ij and TSW (r = 0.756), SOsc (r = 0.612), and Cp (r = 0.337), suggesting the phenological synchronization determined the seed quality and yield. Comprehensively, this study provided fertility and phenology information for management of a Chinese fir seed orchard, as well as a reference for the establishment of advanced seed orchards of conifer trees.

1. Introduction

Seed orchards, which consist of populations with excellent clones or families of economically important trees, are established to produce seeds with high genetic quality [1,2,3]. Higher genetic gains in seed orchards are expected without pollen from unimproved trees, full synchrony, and random mating between clones [4,5]. Therefore, flowering phenology and synchronization and fertility variation are important considerations in the genetic management of seed orchards [6].

Studies on fertility variation are the basis for seed orchard management and for parent selection in advanced-generation seed orchards [7]. The imbalance of gamete contributions in seed orchards affects the genetic composition of the progeny, which can lead to a reduction in genetic diversity [8,9,10,11]. In seed orchards, knowledge about variation in fertility is important for genetic management [8,12]. There are tremendous differences in fertility variation among different tree species, and most previous studies showed that most variation in fertility in forest tree populations (sibling coefficient, Ψ) among trees ranged from 1.00 to 41.67 [13,14]. Great differences in male and female strobilus production has been shown to cause fertility variation between clones in many seed orchards, such as those of Pinus densiflora, P. thunbergia, and P. koraiensis [15]. Knowledge of fertility variation is preferred for parental management in seed orchards because of parental management is relatively easy, is inexpensive, and is not survey intensive [3,16]. Fecundity determines the proportion of female or male gamete contributions and ultimately affects genetic gain. Consequently, on the basis of the results of fertility variation among clones in seed orchards, knowledge of fertility variation can help optimize the genetic compositions of the gamete pool to improve genetic gains, such as thinning or replanting.

Genetic exchange among trees and the genetic composition of progeny in seed orchards are substantially influenced by the flower phenology of trees [15,17]. Complete phenology synchronization among parents is needed to achieve the expected overall genetic gain in seed orchards [18]. Weak synchronization between female receptiveness and pollen dispersal decreases the effective population size, leading to an unbalanced contribution of clones, an unbalanced overall selfing proportion, and an increased risk of contamination in seed orchards [2,18,19]. Data from many studies of seed orchards suggest that asynchronization is a serious problem in seed production (Sp) [2,20,21,22]. For example, the low degree of flowering synchronization between clones indicated great potential for selection and optimization in a Schima superba seed orchard [23]. In a P. patula seed orchard, the overlap of flowering between clones in a seed orchard and natural forest stands was shown to cause pollen contamination [2]. Knowledge of synchronization is considered basic information for pollen management in seed orchards and advanced-generation seed orchard establishment [24]. Therefore, understanding the phenology and synchronization could enable optimization of the configuration of parents, thereby increasing the overall level of outcrossing in seed orchards.

Chinese fir (Cunninghamia lanceolata) is an important, fast-growing timber tree species used for afforestation in southern China [25]. To date, the phenology and pollen distribution in first-generation seed orchards and flowering synchronization between elite trees have been studied comprehensively [9,20,26,27]. Chen et al. (1995) investigated the reproductive phenology of first-generation Chinese fir seed orchards in Guizhou Province and reported that the average synchronization index was very low (0.30–0.37). The current clonal Chinese fir seed orchards were established early and have provided seeds for afforestation, but low Sp and quality still occur in some of the seed orchards. Flowering phenology is influenced by physical factors (e.g., insolation, exposure within seed orchards, and temperature), and plantation establishment of and parental combinations in seed orchards also affect flowering synchronization. Therefore, further quantification of fertility variation and phenological synchronization are fundamental information for seed orchard management to obtain high Sp and quality and are referenced for the establishment of advanced-stage seed orchards. Therefore, the objectives of the study were to (1) determine the fertility variation in seed orchards via flowering data, (2) determine the degree of phenology synchronization, and (3) discuss the effects of fertility and synchronization variation on seed production.

2. Materials and Methods

2.1. Investigated Seed Orchard

A second-generation clonal seed orchard (CSO) of Chinese fir at the Hongya state-owned forest farm in Sichuan, China (29.70° N, 103.26° E; 960–1200 m above sea level), was selected as the survey and study plot. The CSO was established in 2003 through grafting, and the plots were established in accordance with a completely randomized block design, with an isolation belt of Cryptomeria fortunei artificially planted trees. The orchard covers an area of 8.7 ha, with 110 clones in a 3 × 4 m row spacing, and the distance between ramets of the same clone is 20 m; there are 30–66 ramets per clone (mean = 48, median = 41), which were randomly arranged in the blocks. The CSO consists of 12 blocks, and each subline contains all 110 clones, with a mean of 5.3 ± 0.5 ramets per clone. Eighty-seven clones (3–4 ramets per clone, for a total of 305 ramets) were selected for the survey. These 87 clones were selected from seven provenance sites in Sichuan and Chongqing provinces, China: the Yuejiang forest farm (14 clones), Xuyong forest farm (21 clones), Nanchuan forest farm (19 clones), Junlian forest farm (14 clones), Hongya state-own forest farm (7 clones), Fushun forest farm (9 clones), and Gongxian seed orchard (3 clones) (

Table 1. Information of clones in the second-generation clonal seed orchard.

Origin	Longitude (°)	Latitude (°)	Collection No.
Fushun forest farm, Sichuan Province, China	104.85	29.01	EQ48, EQ60, EQ59, EQ61, EQ58, EQ46, EQ41, EQ42, EQ43
Gongxian seed orchard, Sichuan Province, China	104.71	28.45	T91, T72, T73
Hongya state-own forest farm, Sichuan Province, China	103.26	29.70	T209, T204, T205, T206, T217, T218, T164
Nanchuan forest farm, Chongqing Province, China	107.14	29.14	T313, T301, T312, T314, T311, T315, T316, T317, T309, T308, T307, T319, T320, T321, T305, T334, T336, T335, T328
Xuyong forest farm, Sichuan Province, China	105.44	28.16	T358, T359, T367, T361, T360, T372, T363, T378, EQ22, EQ23, T382, T379, EQ24, EQ25, T380, EQ27, T312R, EQ28, EQ21, EQ29, T376
Yuejiang forest farm, Sichuan Province, China	104.52	28.44	T2, T3, T6, T7, T5, T8, T13, T12, T10, T9, T14, T15, T16, T17
Junlian forest farm, Sichuan Province, China	104.50	28.16	T101, T113, T114, T115, T119, T120, T105, T103, T109, T121, T122, T107, T106, T112

). Four branches from the four cardinal directions per tree were marked for the flowering surveys, phenological data collection, and seed surveys.

2.2. Flowering Survey

The production of strobili was investigated by the standard branch method [9]. Four standard branches (lateral branches facing the eastern, southern, western, and northern directions) per individual tree were marked. The number of male (Nm) and female flowers (Nf) on a standard branch were recorded.

2.3. Phenological Data Collection

The phenological progress of male and female strobili was monitored on four marked branches at flowering time (approximately from March to April). These data were recorded every two days (7 March–2 April 2020). The male and female flowers were divided into four stages [9]. The stages of the male flowers were as follows: 1, initial stage (IS_f, staminate strobilus whitening with little pollen shedding); 2, peak stage (PS_f, a large amount of pollen shedding); 3, final stage (FS_f, 90% of pollen has shed); and 4, end stage (ES_f, 100% of pollen has shed, and the staminate strobilus has browned). The stages of the female flowers were as follows: 1, initial stage (IS_m, female flower bud is yellow, with invisible ovules; 2, peak stage (PS_m, at least two whorls of ovuliferous scales expanding, while the ovule is golden yellow with droplets); 3, final stage (FS_m, ovule no longer fresh, with little droplets); and 4, end stage (ES_m, ovuliferous scales completely closing). The days until male and female flowering of each clone were also recorded as the flowering duration of males (FD_m) and females (FD_f), respectively. Julian dates for every stage were recorded for each tree.

The individuals were divided into three types (early, middle, and late) by the beginning of pollen shedding and female receptivity [20]. Early individuals began shedding pollen (M_E) or were receptive (F_E) within 2 days before the earliest individuals, late individuals began shedding pollen (M_L) or were receptive (F_L) within 2 days after the latest individuals, and all other individuals were classified as “middle” in terms of the timing of their pollen shedding (M_M) and receptivity (F_M).

2.4. Evaluation of Seed Production

Cones were removed by hand from the marked branches. Then, the cone production (Cp) of each marked branch was investigated, and the Cp of each tree was calculated. Then, the cones were dried, and all cones from each tree were randomly selected to measure the seed production (Sp, g) and seed output of a single cone (SOsc, g). The thousand-seed weight (TSW, g) per tree was also measured.

2.5. Data Analysis

The variance of fertility, phenology, and production were calculated according to the general linear model (GLM) procedure by SPSS software according to the following statistical model [28]. “Clone” was nested within “Provenance” for all analyses of variance (ANOVAs):

$$Y_{ij}=\mu +P_i+{\delta }_{ij}+e_{ij}$$

(1)

where $Y_{ij}$ is the fertility, phenology, or production of clone j; $\mu$ is the overall mean; P_i is the effect of provenance; ${\delta }_{ij}$ is the effect of the clone in which the clone is nested in the provenance; and $e_{ijk}$ represents the random error.

The clonal broad-sense heritability ($H_c^2$) for fertility (flower production), phenology, and seed traits was calculated as follows [29]:

$$H_c^2=\frac{{\sigma }_c^2}{{\sigma }_{total}^2}$$

(2)

where ${\sigma }_c^2$ is the estimate of variance for clones and ${\sigma }_{total}^2$ is the total estimate of the variance, which includes both genetic and environmental variance.

The coefficients of variation (CV_f for the female strobilus; CV_m for the male strobilus) were calculated and applied to estimate the fertility variation. The variation in female (ψ_f) and male fertility (ψ_m) among individuals was estimated as follows [15]:

$${\psi }_f=C{V}_f^2+1$$

(3)

$${\psi }_m=C{V}_m^2+1$$

(4)

The correlation between female and male strobilus production (r) was analyzed via Pearson’s correlation coefficients. The sibling coefficient (ψ), which is derived from combined fertility variation, is calculated by the following formula if there is a significant correlation between female and male strobilus production [30].

$$\psi =0.25\left({\psi }_f+{\psi }_m\right)+0.5\left[1+r\times \sqrt{\left({\psi }_f-1\right)\left({\psi }_m-1\right)}\right]$$

(5)

If no correlation is detected between female and male strobilus production, the ψ value is calculated as follows [16]:

$$\psi =0.25\left({\psi }_f+{\psi }_m\right)+0.5$$

(6)

The parameters of effective number of parents (N_p) in the seed orchard is calculated as N_p = n/ψ, where n is the total number of ramets in the seed orchard (the same as below). Similarly, the effective number of female (N_pf = n/ψ_f) and male gametic parents (N_pm = n/ψ_m), the relative effective number of parents (N_r = N_p/n), and the relative effective numbers of female (N_rf = N_pf /n) and male parents (N_rm = N_pm /n) were also calculated [31,32,33]. A parent–balance curve was used to estimate the female and male gametic contributions. The trees were ranked from high to low according to their female and male strobilus production, and then the cumulative percentage calculations were plotted against the equal cumulative percentage in the seed orchard [34].

There were significant differences in the amount of pollen shed and the degree of receptivity between each phenological stage. A weight value was assigned to each phenological stage to fully reflect the differences between clones. The weight values of female flowers in the IS_f, PS_f, and FS_f were 0.4, 1.0, and 0.2, respectively, and the weight values of male flowers in the IS_m, PS_m, and FS_m were 0.2, 1.0, and 0.2, respectively [9]. The proportion of pollen shed or receptivity equaled the sum of the frequency multiplied by the weight value at each stage in a whole day. The phenological synchronization index (PO_ij) was then calculated as follows [35]:

$$P{O}_{ij}={{\displaystyle \sum }}_{k=1}^n\min (M_{ki},P_{kj})/{{\displaystyle \sum }}_{k=1}^n\max \left(M_{ki},P_{kj}\right)$$

(7)

where PO_ij is the phenological synchronization index between the ith and jth clones, M_ki is the proportion of pollen shedding of the ith clone on day k, P_kj is the proportion of the female receptivity of the jth clone on day k, and n is the number of days of total flower duration of i and j. The variance of the synchronization index among different types (early, middle, and late) of female and male flowering was calculated by one-way ANOVA.

Linear regression analysis between the IS_f and IS_m and the FD_f and FD_m was performed to determine the influence of the initial stage of male and female flowering on flowering duration. Moreover, correlations between diameter at breast height (DBH) and N_f, N_m, IS_f, IS_m, and female and male flower duration (FD_f, FD_m) were analyzed by linear regression analysis. Linear regression analysis between PO_ij and C_p, Sp, SO_sc, and TSW was also performed to determine the implications of phenology synchronization concerning Sp and quality.

3. Results

3.1. Fertility Variation in the CSO

In the current study, the obvious variation of male and female flower production among clones occurred in the Chinese fir CSO (

Table 2. GLM-based ANOVA for flower number in the CSO.

Source	Traits	df	Type III SS	Mean Square	F	Hc2
Provenance	Nm	6	37,855.434	6309.239	0.230
	Nf	6	3618.721	603.120	1.750
Clone (provenance)	Nm	79	1,770,227.261	22,407.940	0.808	0.396
	Nf	79	35,415.797	448.301	1.285	0.318
Error	Nm	55	1,507,138.583	27,402.520
	Nf	55	18,952.333	344.588

Note: N_m, number of male flowers; N_f, number of female flowers. A total of 80 and 56 clones with male and female flowers, respectively, were included in the investigation; therefore, male flower data were collected from 80 clones, and female flowers data were collected from 56 clones.

). A large difference was detected between male and female strobilus production, which was also the case among clones. The Nf and Nm varied among individuals within the seed orchard, and fertility was low (ψ = 2.03). The parent–balance curves for Nm and Nf revealed a significant deviation from balanced production (Figure 1). The most abundant 10 tree types (approximately 7% of total) produced 30.47% of Nm, and the most abundant of five tree types (approximately 6% of total) produced 32.63% Nf, indicating that the female and male parents unequally contributed to the gamete gene pool in the seed orchard. The coefficients of variation of females (ψ_f = 2.85) and males (ψ_m = 2.69) were low among clones. There was small difference in N_p (93.59); the N_pf (66.61) was lower than the N_pm (0.59). Moreover, N_r, N_rf, and N_rm were low in the seed orchard, with values of 0.49, 0.35, and 0.37, respectively.

3.2. Genetic Variation in Reproductive Phenology

The IS_m and IS_f occurred on 7 March and 9 March, respectively (Figure 2). The IS_m lasted 2–8 days, with an average of 4 days, and the PS_m lasted 2–14 days; afterward, the trees entered the FS_m after an average of 9 days. The IS_f lasted 2–8 days, with an average of 4 days, and the PS_f lasted 4–14 days; afterward, the trees entered the FS_f after 9 days. There was a significantly negative correlation between FD_f and IS_f, with a high R² value (0.464), but there was no correlation between IS_m and FD_m (Figure 3). Clones of the early phenological type shed pollen and sustained receptivity longer than did the clones of the late phenological type. The variation in IS_m and FD_m was significantly influenced by genotype (

Table 3. GLM-based ANOVA for the initial stage and flower duration.

Source	Traits	df	Type III SS	Mean Square	F	Hc2
Provenance	ISm	6	10.056	1.676	1.398
	FDm	6	57.599	9.600	2.040
	ISf	6	2.739	0.457	0.132
	FDf	6	25.332	4.222	0.300
Clone (provenance)	ISm	79	173.713	2.199	1.834 *	0.433
	FDm	79	1025.777	12.985	2.760 **	0.476
	ISf	55	80.768	1.469	0.424	0.272
	FDf	55	556.817	10.124	0.720	0.356
Error	ISm	52	62.333	1.199
	FDm	52	244.667	4.705
	ISf	20	69.333	3.467
	FDf	20	281.333	14.067

Note: IS_m, male flower initial stage; FD_m, male flower duration; IS_f, female flower initial stage; FD_f, female flower duration; * and ** mean significant differences at the levels of 0.05 and 0.01, respectively. A total of 80 and 56 clones with male and female flowers, respectively, were included in the investigation; therefore, male flower data were collected from 80 clones, and female flowers data were collected from 56 clones.

). All phenological traits exhibited moderate-to-low clonal heritability (0.272–0.476), and female receptivity reflected lower genetic control than did pollen shedding.

3.3. Phenological Synchronization

The rate of pollen shedding affects the female receptivity rate during the flowering period, which is beneficial to pollen dispersal (Figure 4B). There was an obvious difference in phenological synchronization among clones/individuals (Figure 4A,C). The phenological synchronicity averaged 0.286, indicating that there was 28.6% overlap between pollen shedding and female receptivity. The synchronization indexes of the clones ranged from 0.000 to 0.585 (with the clone as the pollen donors) and 0.000 to 0.385 (with the clones as the pollen receptors). Our analyses revealed only 28.6% overlap between pollen shedding and female receptivity, which ultimately affected the gamete contribution of the parents and the genetic composition of seeds. Approximately 41.61% of the combinations had a PO_ij of zero, indicating that only approximately 60% of trees had an overlap of female and male flowering. Trees in the early (E), middle (M), and late (L) phenology classes showed significantly different PO_ij values (

Table 4. Average synchronization index at different flowering types.

Male	Female
	Early Phenological Type	Middle Phenological Type	Late Phenological Type
Early phenological type	0.197 a	0.354 c	0.266 b
Middle phenological type	0.489 e	0.461 e	0.333 c
Late phenological type	0.164 a	0.381 d	0.288 b

Note: The same letters mean no significant differences; the different small letters in the table mean significant differences.

). The group comprising F_M and M_M had the highest PO_ij. The M phenology group had a higher mean PO_ij than the other groups did, such as the F_M × M_M and F_E × M_M groups, whose PO_ij values were 0.461 and 0.489, respectively, which were significantly higher than those of the other combinations. The average PO_ij of self-pollination was 0.298, which was higher than that of cross-pollination (0.236), indicating that there was a larger probability of selfing in the seed orchard.

3.4. Relationships between Seed Production, Fertility, and Synchronization

There were no significant differences in Cp or Sp among provenances and clones, but the variation in TSW was significantly influenced by provenance and genotype (

Table 5. GLM-based ANOVA for Cp and Sp.

Source	Trait	df	Type III SS	Mean Square	F	Hc2
Provenance	Cp	5	1539.855	307.971	0.708
	Sp	5	365.211	73.042	2.383
	TSW	5	13.747	2.749	3.740 *
	SOsc	5	0.437	0.087	2.548
Clone (provenance)	Cp	40	18,319.518	457.988	1.053	0.381
	Sp	40	2410.360	60.259	1.966	0.368
	TSW	40	127.04	3.176	4.320 **	0.477
	SOsc	40	1.891	0.047	1.377	0.280
Error	Cp	18	7825.833	434.769
	Sp	14	429.149	30.653
	TSW	13	9.557	0.735
	SOsc	14	0.481	0.034

Note: Cp, cone production; Sp, seed production; SOsc, seed production of a single cone; TSW, thousand-seed weight. * and ** mean significant differences at the levels of 0.05 and 0.01, respectively. Cones and seeds from only 41 clones and 6 provenances were harvested for GLM-based analysis.

). The low value of heritability (0.280–0.477) indicated that the seed traits underwent low genetic control in the seed orchard. The results of linear regression analysis showed that there were significant relationships between PO_ij and both Cp and Sp (Figure 5), but PO_ij showed a weak correlation with both TSW (R² = 0.048, p = 0.280) and SOsc (R² = 0.001, p = 0.868), indicating the phenological synchronization among clones had a significant influence on seed yield.

4. Discussion

Expected genetic gains may not be obtained if the gamete contribution deviates from that of random mating in seed orchards [33,36]. Differences in strobilus production have also been found among individuals or clones within a population of other forest tree species such as Cedrus libani [16], Eucalyptus camaldulensis [33], and Pinus koraiensis [15]. Notably, parent selection is important for Sp and orchard management. The variation in fertility causes gametic phase disequilibrium, which increases the general probability of inbreeding or pollen contamination in seed orchards [9]. The N_p is used to characterize seeds produced from seed orchards; this parameter is based on correlations between male and female fertility. When the N_p is small, the genetic composition of seeds produced in the seed orchard may be different from the expected value [15]. The N_r is used to compare the total number of clones and effective number in the seed orchard [15]. In this study, the N_r (0.49) in seed orchard was low, indicating that as much as 49% of the clones had the same coancestry and an average degree relatedness within the gene pool among seeds in the seed orchard. These findings also implied that a low effective population size leads to a low genetic diversity of seeds. Therefore, balancing the N_p should be fully considered for orchard management. ψ expresses how fertility varies among parents, as this parameter reflects the increase in the probability that sibs occur compared to that when parents have equal fertility [31]. As in previous reports of C. libani seed orchards [16], the ψ in the seed orchard in the current study was 2.03, indicating that the probability that two individuals share a parent was twice as high as that when the parental fertility is equal across the seed orchard [13,16]. Hence, knowledge of the gamete contribution to variation in seed orchards is very valuable when breeders wish to obtain the maximum expected genetic gain.

Phenological synchronization is the basic information of major importance for management in seed orchards [24]. In our study, the genetic differences among clones for IS_m and FD_m was reflected by a GLM. As in other seed orchards [17,24], significant variation in flowering phenology among clones occurred in the Chinese fir seed orchard in the current study. However, unlike that of Chinese fir clones at another site, the male flowering of clones in the seed orchard in the present study started approximately 2 days before the female flowering [20]. The pollen shedding stage occurred prior to female receptivity, which was beneficial to pollination in the seed orchard. Regression analysis of the initial stage against pollen shedding and receptivity duration indicated that the pollen shedding and receptivity of the early flowering clones were significantly longer than those of the late-flowering clones. This suggested that duration of flowering is influenced by the initiation of flowering when the climate, temperature, and light conditions are consistent. Low synchronization results in nonrandom mating in seed orchards, reducing the genetic gain and diversity of seeds [24,37]. Somewhat lower phenological synchronization was also detected in a Quercus rubra seed orchard (PO_ij = 0.30) [17], first-generation Chinese fir seed orchard (PO_ij = 0.30–0.37) [9], and P. tabuliformis seed orchard (PO_ij = 0.12–0.38) [24]. Because flowering differences resulted in decreased PO_ij values, in the current study, the low PO_ij of early flowering types of female flowers and of the early and late-flowering types of male flowers were likely caused by lower PO_ij values across orchards. In our study, the average PO_ij of the female parents was slightly higher than that of the male parents, but the N_f was significantly lower than the N_m. These results indicated that female flower management is more important for Sp than male flower management is under the premise that the seed orchard can still produce enough pollen and is a major factor affecting the genetic gain of seeds in seed orchards of Chinese fir. Moreover, self-pollination reduces the genetic gain and diversity of seeds. As in previous reports on first-generation seed orchards of Chinese fir [9], in this study, most clones had a high risk of self-pollination, as predicted by the higher average PO_ij of self-pollination compared with cross-pollination. Measures should be utilized to decrease these effects, such as increasing the number of clones and selecting dichogamy parents.

The difference in N_f and N_m and flowering asynchrony cause imbalances in gamete contribution, decreased genetic diversity, and a high selfing rate and pollen contamination rate in seed orchards [9]. To achieve maximum seed yields, all female strobili must be pollinated by the pollen of male flowers in the seed orchard when the former are fully receptive to pollen [16]. The flowering habit of clones have a substantial effect on the genetic composition of seeds in seed orchards. Seed germination time, germination rate, and germination potential significantly influenced the TSW of Chinese fir [38]. Positive significant correlations were found between PO_ij and seed characteristics in the seed orchard, indicating that the final seed yield and quality are influenced by phenological synchronization among clones. Consequently, this study is essential for providing basic information for making decisions for seed orchard management and future establishment of advanced-stage seed orchards to obtain the maximum expected genetic gain; these decisions include practices such as pruning, increasing biomass, controlling pollination, and adjusting parent configurations.

5. Conclusions

Significant variation in fertility and flowering phenology occurred between clones in a second-generation seed orchard of Chinese fir. The most abundant types of trees (7% and 6%) produced 30.47% and 32.63% male and female flowers, respectively, and the gamete contributions of the female and male flowers were unbalanced among clones in the seed orchard. A low synchronization index (PO_ij = 0.298) and high risk of selfing in the seed orchard indicate that the rate of pollen shedding affects female receptivity during flowering. The average PO_ij of female parents was slightly higher than that of male parents, but the N_f was significantly lower than the N_m; therefore, female flower management is more important for Sp than male flower management under the premise that the Chinese fir seed orchard could produce enough pollen needed in the future. Genetically, there is a low clonal H_c² (0.272–0.477) for fertility and flower production. We also found a significantly positive correlation between PO_ij and Sp and Cp, indicating that phenological synchronization affects seed quality and yield. Overall, management practices such as increasing biomass, controlling pollination, and resetting of parents are necessary to improve flowering synchronization and the balance of gamete contributions, which are beneficial for increasing Sp and quality in Chinese fir seed orchards. Furthermore, continuous monitoring of flowering phenology in seed orchards is important for obtaining high genetic gains and for providing reference data for the establishment of advanced-stage seed orchards.