Tier-specific physiological profiling in male rowers: insights from an integrated three-phase performance diagnostic framework

Abstract

Background

This study aimed to characterize the tier-specific physiological signatures of male rowers using an integrated three-phase performance diagnostic protocol, delineating differences in anaerobic power, aerobic characteristics, and power maintenance capacity across competitive levels.

Methods

Fifty-one male single scull rowers were assigned to three tiers (Elite, Sub-elite, and Developmental) according to China’s official athlete classification standards. All participants completed a standardized testing protocol on a rowing ergometer, which included a 30-second all-out anaerobic test, an incremental aerobic test, and a 120-second all-out test to assess the capacity to sustain power under fatigue.

Results

MANCOVA revealed a significant multivariate effect of competitive tier on anaerobic performance (F = 3.03, p = 0.002, η²=0.32), aerobic performance (F = 8.53, p < 0.001, η²=0.72), and performance in the 120-second test (F = 5.53, p < 0.001, η²=0.39). Elite rowers exhibited superior anaerobic power, with significantly higher peak and mean power, lower fatigue index, and elevated peak blood lactate in the 30-second test (all p < 0.05). During incremental testing, they demonstrated greater exhaustive load, Power output at V̇O₂max (W), and GET/V̇O₂max ratio (all p < 0.001), collectively pointing to a superior integrated aerobic profile, characterized by higher maximal output and threshold. In the 120-second test, Elite athletes maintained higher power output and stroke rate despite similar peak blood lactate levels, highlighting superior short-term power maintenance and metabolic regulation.

Conclusions

The findings establish a discriminative, tier-specific physiological profile in rowing performance. Elite athletes are distinguished by an integrated superiority in anaerobic power, aerobic capacity, and short-term power maintenance under fatigue. The three-phase diagnostic framework provides a practical and discriminative tool for moving beyond result-based classification, enabling targeted, physiology-guided training interventions.

Introduction

Rowing competitions are standardized with a fixed race distance of 2000 m. This high-intensity, long-duration event rigorously tests athletes’ physical endurance, demanding exceptional aerobic capacity, maximal anaerobic power output, and a high tolerance to elevated lactate levels [1].From an exercise physiology perspective, the race progression imposes distinct, sequential, and phase-specific demands on the energy systems.Substantial empirical evidence indicates that elite rowers consistently adopt a “reverse J-shaped” pacing strategy during 2000-meter races, characterized by a fast start, a sustained high-speed middle segment, and an all-out final sprint (i.e., a fast-slow-slow-fast pattern) [2, 3]. This widely observed pacing phenomenon underscores fundamental differences in the dominant physiological mechanisms across race phases: The Launch Phase requires athletes to achieve peak power output to establish an initial speed advantage [4, 5]. This phase depends critically on explosive energy provision from the phosphagen system and a rapid glycolytic response, demanding both a high rate of force development and the ability to rapidly recruit high-threshold motor units [6]. The Mid-Phase (constituting the majority of the race) demands the maintenance of a high yet sustainable power output. Performance in this phase is primarily dependent on aerobic metabolism and is determined by an integrated profile of key physiological traits: a high maximal oxygen uptake (V̇O₂max) defining the aerobic power ceiling, superior exercise economy minimizing the oxygen cost of a given workload, and a high gas exchange threshold (GET) indicating the ability to sustain a large fraction of V̇O₂max without rapid lactate accumulation [7]. The Sprint Phase (the final segment of the race) demands a significant increase in power output amidst high levels of cumulative fatigue and metabolic acidosis. Success in this phase hinges on the capacity to maintain power output under severe physiological duress [8].

The current athlete classification system in rowing, both internationally and within frameworks such as China’s official standards [9], primarily relies on final competition rankings. This result-oriented approach, while pragmatic, fails to quantify the specific physiological traits an athlete exhibits during distinct segments of the race. For instance, two rowers with similar overall 2000-meter times may demonstrate significant differences in their starting power or their capacity to sustain power in the final sprint. Conventional assessment methods like the 2000-meter ergometer time trial effectively measure global performance but are fundamentally limited in their ability to decouple the phase-specific functions of different energy systems (e.g., anaerobic power at the start, aerobic efficiency during the mid-race, sprint power decay resistance) [10]. This lack of diagnostic granularity hampers the development of precisely targeted physiological interventions. As underscored in the broader sports science literature, moving beyond holistic measures to assess sport-specific energetic capacities is crucial for effective training guidance [11, 12]. Therefore, employing these official, result-based tiers as the grouping variable establishes an objective and ecologically valid framework to investigate a central question: do athletes separated by competition outcomes also demonstrate systematic, graded differences in the physiological capacities that underpin race performance?

Consequently, the primary aim and applied value of this study lie in developing and implementing an integrated three-phase diagnostic framework that directly corresponds to the physiological phases of the race. This framework comprises: (1) a 30-second all-out test to assess anaerobic power relevant to the start phase; (2) an incremental exercise test to evaluate aerobic capacity and metabolic thresholds crucial for the middle phase; and (3) a fatigue-primed 120-second all-out test to simulate and assess power maintenance capability indicative of the final sprint phase. By systematically comparing the performance of Elite, Sub-elite, and Developmental level athletes within this framework, this study aims to establish a diagnostic tool that moves beyond ranking-based classification. Its goal is to precisely identify physiological limitations and provide a direct basis for individualized training prescription.

Methods

Participant

Fifty-one healthy, non-smoking male single scull rowers of East Asian ethnicity were stratified into three competitive tiers according to China’s official athlete classification system—the Technical Level Management Measures for Athletes (State General Administration of Sports, 2024) [9]:(1) Elite Tier (n = 17): National Elite Athlete or International Elite Athlete, certified by winning medals in national-level events (e.g., National Games) or international competitions. (2) Sub-Elite Tier (n = 17): Class I Athletes, certified by achieving performance standards in provincial championships or national U-series regattas. (3) Developmental Tier (n = 17): Class II Athletes, certified by meeting standards in provincial youth competitions. Inclusion criteria required participants to be injury-free for ≥ 1 month preceding testing and in good general health. The experiment was conducted at the end of the preparation phase, when the athletes had developed a solid fitness base and were free from competitive commitments, allowing for better control of external variables. Subjects abstained from alcohol and caffeine-containing beverages for ≥ 48 h and avoided high-intensity lower-limb exercise for ≥ 24 h pre-testing. Written informed consent was obtained after full disclosure of study objectives, procedures, benefits, and risks. The experimental protocol was approved by the Institutional Review Board of Jiangxi Normal University (IRB-JXNU-PEC-2024019), and all procedures complied with the ethical principles outlined in the Declaration of Helsinki. Basic information is shown in Table 1.

Table 1.

Basic information of participants (N = 51)

Group	N	Height(cm)	Weight(kg)	Age (years)	Training years (years )
Elite Tier	17	190.2 ± 5.0	84.8 ± 6.1	20.0 ± 1.8	4.8 ± 1.2
Sub-Elite Tier	17	189.1 ± 4.4	84.8 ± 7.7	19.3 ± 1.3	4.5 ± 1.2
Developmental Tier	17	189.2 ± 5.3	82.5 ± 9.7	17.1 ± 1.0	2.9 ± 0.9

Pre-test process and standardized control

Prior to formal testing, all participants completed a minimum of two familiarization sessions with the full testing protocol on separate days to mitigate learning effects and ensure performance reflected true physiological capacity. Pre-test surveys were administered to document participant ethnicity (all self-identified as Han Chinese), training years, and competition level.

All testing was conducted within a single week in the same laboratory, within an environmental chamber with conditions maintained at 22 ± 2 °C and 50 ± 10% relative humidity. Testing was scheduled during the general preparation phase of the annual training plan, a period of standardized training, to minimize variations in fitness and fatigue levels. All tests were conducted during daytime hours (between 09:00 and 18:00).

Baseline anthropometric data (height, body mass) were recorded. Pre-exercise blood lactate (BLapre) was measured prior to the warm-up protocol. The Jaeger Oxycon mobile cardiopulmonary function monitor (Oxycon Mobile) was initialized and calibrated one hour prior to data collection. Calibration procedures followed manufacturer specifications, including gas analyzer calibration using certified precision gases (16.0% O₂, 5.0% CO₂, balanced with N₂) and automated flow calibration of the integrated Triple V volume transducer. All rowing tests were performed on a Concept2 Model D rowing ergometer.

Warm-up protocol

Following baseline assessments, athletes performed a standardized 13-minute warm-up on a rowing ergometer (Model D, Concept2, Inc., USA). This ergometer has established validity and reliability for measuring rowing performance. The warm-up began with a 10-minute aerobic graduation phase, comprising 3 min at 60%, 3 min at 70%, and 4 min at 80% of age-predicted maximum heart rate. Immediately thereafter, a 3-minute neuromuscular activation phase was performed, consisting of three 10-second all-out sprints, each separated by 50 s of passive rest or very light paddling. Upon completion of the warm-up, a strict 10-minute recovery period was observed before the commencement of any maximal exercise tests. Throughout the warm-up and all subsequent testing, the damper was adjusted to maintain a standardized drag factor of 130 for all participants, which was verified and recorded prior to each session.

Testing protocol

The formal testing battery was conducted on a rowing ergometer and comprised three discrete tests (Fig. 1). Breath-by-breath data were collected throughout using a pre-calibrated Oxycon Mobile cardiopulmonary function monitor. Raw ventilatory data were processed by the system’s software to automatically exclude aberrant breaths (e.g., from coughing or swallowing), with final values exported as 5-second averages for analysis.

Fig. 1.

Flowchart of experimental design. Incremental Rowing Test:Initial working, body weight multiplied by 2.0, An increment of 30 W every 2 min, at a constant speed rowing for 2 min

The protocol began with a 30-second all-out rowing test to assess maximal anaerobic power and glycolytic contribution [13]. Athletes were required to start with a maximal sprint from the first stroke and maintain this all-out pacing for the entire duration. Peak power (PP), mean power (MP), fatigue index (FI), and mean stroke rate were recorded. Capillary blood samples (10 µL) were taken from the earlobe at the 3rd and 7th minutes of recovery, with the highest value defined as the BLapeak. Subsequently, aerobic power was evaluated using an incremental test protocol, starting at a power output of 3.0 × body mass (kg), with increments of 30 W every 2 min until volitional exhaustion. This load-increment scheme was designed to elicit exhaustion within 8–12 min, aligning with established recommendations for maximal cardiopulmonary exercise testing [14] and ensuring a sufficient number of stages for robust determination of metabolic thresholds in national-level rowers. Throughout the test, athletes were required to maintain the target power within a ± 5 W range, enforced via real-time monitoring and standardized verbal feedback from the test conductors. The test was terminated if an athlete was unable to sustain the required power for 10 consecutive seconds, ensuring consistent power output management. Recorded metrics included the number of completed stages and gas exchange parameters: V̇O₂max, PV̇O₂max, V̇O₂max_PD (defined as a plateau in oxygen consumption with a change of ≤ 50 ml/min) [15], V̇Emax, and GET (determined by the V-slope method as the non-linear increase in VCO₂ relative to VO₂) [16, 17]. Finally, to evaluate fatigue resistance—defined as the ability to sustain high-intensity performance under cumulative fatigue—a 120-second all-out rowing test was administered. Participants were instructed to maintain a maximal sprint from the outset, with the test terminating if their power output fell below their individually predetermined power at V̇O₂max (PV̇O₂max) for 5 consecutive seconds. Standardized verbal encouragement was provided to ensure maximal effort. PP, MP, FI, and stroke rate were recorded to quantify performance under this high-stress paradigm.

Post-exercise earlobe blood samples (10 µL) were collected at the 3rd and 5th minutes of recovery. These specific time points were selected based on preliminary data and serial lactate measurements from our laboratory, which identified them as the most frequent instances of peak blood lactate concentration following all-out rowing efforts in this athlete population. The highest value obtained from these two time points was recorded as BLApeak. Expired gas data were collected breath-by-breath throughout the test using a metabolic cart. Blood lactate concentration was determined using an EKF C-Line Clinic Blood Lactate Analyzer (Germany). Standardized verbal encouragement was provided to ensure maximal effort. Fatigue Index (FI) was calculated as the percentage decline in power from the peak to the minimum value during the test, using the formula: FI (%) = [(P_peak – P_min) / Ppeak] × 100% [18]. Key parameters are summarized in Table 2.

Table 2.

Core metrics for sport-special capacity testing in rowers

Testing stage	Core Metrics
Anaerobic Power Test (30-sec All-out Rowing Test)	Peak power output (PP), Mean Power Output (MP), Fatigue Index (FI), Stroke Rate, Pre-exercise Blood Lactate (BLapre), Peak Blood Lactate (BLapeak).
Aerobic Capacity Test (Incremental Rowing Test)	Initial load, Exhaustion Load, number of workload increments,, Maximum Oxygen Uptake (V̇O2max), Relative Maximum Oxygen Uptake (V̇O2max/kg), Power output at Maximum Oxygen Uptake (PV̇O2max), Plateau Duration at Maximum Oxygen Uptake (V̇O2maxPD), Gas Exchange Threshold (GET), Proportion of Gas Exchange Threshold to Maximum Oxygen Uptake (GET/V̇O2max), Maximum Minute Ventilatory (V̇Emax)
Fatigue Resistance Test (120-sec All-out Rowing Test)	Peak Power Output (PP), Mean Power Output (MP), Fatigue Index (FI), Stroke Rate, Peak Blood Lactate (BLapeak).

Participant pre-test controls

To minimize confounding variables, participants were instructed within the 48 h preceding the test to: avoid high-intensity exercise, abstain from caffeine and alcohol, and maintain their habitual nutrition and sleep routines. On the test day, they were required to consume a standardized meal at least two hours before arrival and to ensure adequate hydration.

The research assistants conducting the tests were aware of the study’s general aims. However, all subsequent data analysis was performed by a researcher who was blinded to the competitive tier grouping of the participants to prevent ascertainment bias.

Statistical analysis

Data are presented as mean ± standard deviation. Statistical analyses were performed using SPSS 25.0. The normality of data distribution was assessed using the Shapiro-Wilk test. An a priori sample size estimation was conducted using G*Power 3.1.9.7 for a MANCOVA comparing three groups, with four covariates. Based on an anticipated large effect (f²(V) = 0.25, corresponding to ηp² ≈ 0.20), which is plausible when comparing athletes across distinct performance tiers, and given α = 0.05, and power = 0.80, the analysis recommended a total sample size of N = 51. The final study sample comprised 51 elite rowers (17 per tier). To assess the detectable effect size of the achieved sample, a sensitivity analysis was performed. This analysis established that the present design (N = 51, α = 0.05, power = 0.80) is adequately powered to detect a large effect size of f²(V) ≥ 0.25 (corresponding to ηp² ≥ 0.20). Statistical inferences were based on MANCOVA, with Bonferroni-adjusted post-hoc comparisons used to delineate specific between-tier differences. Effect sizes were reported to quantify the magnitude of differences: η² was interpreted as small (0.01–0.05), medium (0.06–0.13), and large (≥ 0.14). Statistical significance was defined as p < 0.05.

Results

Anaerobic performance disparities across competitive tiers in male rowers

A multivariate analysis of covariance (MANCOVA) was conducted to examine differences in anaerobic performance across competitive tiers while controlling for covariates. The analysis revealed a significant main effect of group ( F = 3.03, p = 0.002, η² = 0.32), indicating substantial differences in the combined anaerobic performance measures among Elite, Sub-Elite, and Developmental athletes. Subsequent univariate analyses, after adjusting for weight, height, age, and training years, identified significant group effects on key performance metrics .

Post-hoc comparisons with Bonferroni correction elucidated the specific between-tier disparities (see Tables 3 and 4 ). Regarding PP, Elite athletes (787.94 ± 76.32 W) produced significantly higher values than Developmental athletes (705.35 ± 89.46 W) (F = 6.80, p = 0.037, η² = 0.13). For MP, Elite athletes (713.76 ± 53.81 W) produced significantly higher values than both Sub-Elite (673.06 ± 55.45 W) (F = 5.41, p = 0.07, η² = 0.11) and Developmental athletes (636.12 ± 83.39 W) (F = 7.91, p = 0.022, η² = 0.15). Elite athletes exhibited a significantly lower fatigue index (FI) (12.0 ± 3.7%) than both Sub-Elite (15.4 ± 3.5%) (F = 8.51, p = 0.017, η² = 0.16) and Developmental tiers (17.7 ± 2.8%) (F = 16.50, p = 0.001, η² = 0.27). Furthermore, Elite athletes generated significantly higher BLapeak levels (8.31 ± 0.91 mmol/L) compared to both Sub-Elite (7.05 ± 1.75 ) (F = 7.19, p = 0.031, η² = 0.141) and Developmental athletes (6.75 ± 1.47 mmol/L) (F = 7.82, p = 0.023, η² = 0.15). In contrast, no significant group differences were detected for stroke rate or pre-test blood lactate levels (all p > 0.05).

Table 3.

Comparisons of anaerobic performance and stroke rate during a 30-s All-Out rowing test among rowers of different competitive tiers

Variables	Elite Tier	Sub-Elite Tier	Developmental Tier
PP (W)	787.94 ± 76.32#	740.82 ± 63.83	705.35 ± 89.46
MP (W)	713.76 ± 53.81*#	673.06 ± 55.45	636.12 ± 83.39
FI (%)	12.0 ± 3.7*##	15.4 ± 3.5	17.7 ± 2.8
Stroke rate (spm)	55.9 ± 5.3	53.0 ± 4.6	52.1 ± 6.3
BLapre (mmol/L)	1.89 ± 0.39	1.76 ± 0.34	1.63 ± 0.41
BLapeak (mmol/L)	8.31 ± 0.91*#	7.05 ± 1.75	6.75 ± 1.47

*p < 0.05, **p < 0.01 vs. Sub-Elite Tier ; #p < 0.05, ##p < 0.01 vs. Developmental Tier

Table 4.

MANCOVA of anaerobic performance and stroke rate during a 30-s all-out rowing test by competitive tier

Variables	General F of three tiers	Group effect	Between-group effect
			Elite Tier vs.Sub-Elite Tier F, P, η²	Elite Tier vs.Developmental Tier F, P, η²	Sub-Elite Tier vs.Developmental Tier F, P, η²
PP (W)	F = 3.03, P = 0.002, η²=0.32	F = 4.00, P = 0.025 ,η²=0.15	F = 5.02, P = 0.090, η²=0.10	F = 6.80, P = 0.037, η²=0.13	F = 1.02, P = 0.956, η²=0.023
MP (W)		F = 4.54, P = 0.016 ,η²=0.17	F = 5.41, P = 0.07, η²=0.11	F = 7.91 P = 0.022, η²=0.15	F = 1.36, P = 0.752, η²=0.03
FI (%)		F = 8.67, P = 0.001 ,η²=0.28	F = 8.51, P = 0.017, η²=0.16	F = 16.50, P = 0.001, η²=0.27	F = 3.72, P = 0.153, η²=0.08
Stroke rate (spm)		F = 0.92, P = 0.406 ,η²=0.04	F = 1.62, P = 0.628, η²=0.04	F = 1.07, P = 0.923, η²=0.02	F = 0.01, P = 1.000, η²<0.01
BLapre (mmol/L)		F = 1.30, P = 0.284 ,η²=0.06	F = 1.26, P = 0.800, η²=0.03	F = 2.42, P = 0.380, η²=0.05	F = 0.61, P = 1.000, η²=0.01
BLapeak (mmol/L)		F = 5.03, P = 0.011,η²=0.19	F = 7.19, P = 0.031, η²=0.141	F = 7.82, P = 0.023, η²=0.15	F = 0.73, P = 1.000, η²=0.02

Among the covariates, only body weight significantly predicted PP and MP. The results establish a clear hierarchy in anaerobic capabilities, with the most pronounced advantages observed in Elite athletes across multiple power output and metabolic response indicators.

Aerobic performance disparities across competitive tiers in male rowers

Aerobic performance was assessed via an incremental test. After controlling for body weight, height, age, and training years, a MANCOVA revealed a significant main effect of competitive tier on the combined aerobic performance measures ( F = 8.53, p < 0.001, η² = 0.72), indicating substantial differences in aerobic capacity across tiers. Subsequent univariate analyses confirmed a clear hierarchical gradient (Elite > Sub-Elite > Developmental) for exhaustive load, the number of workload increments,, power at VO₂max, and the GET/VO₂max ratio.

Bonferroni-adjusted post-hoc comparisons detailed this gradient (see Tables 5 and 6). Elite athletes achieved a significantly higher exhaustive load (401.41 ± 23.00 W) than both Sub-Elite (367.24 ± 20.82 W; F = 27.42, p < 0.001, η² = 0.40) and Developmental athletes (328.88 ± 38.19 W; F = 68.84, p < 0.001, η² = 0.62). They also completed more workload increments (5.8 ± 0.5) than Sub-Elite (4.8 ± 0.7; F = 28.22, p < 0.001, η² = 0.40) and Developmental athletes (3.7 ± 0.8; F = 80.37, p < 0.001, η² = 0.66), generated greater power at VO₂max (401.41 ± 23.00 W) than Sub-Elite (370.00 ± 24.90 W; F = 22.33, p < 0.001, η² = 0.35) and Developmental athletes (330.35 ± 33.55 W; F = 62.00, p < 0.001, η² = 0.60), and exhibited a higher GET/VO₂max ratio (85.32 ± 3.93%) than Sub-Elite (77.36 ± 2.72%; F = 93.45, p < 0.001, η² = 0.69) and Developmental athletes (69.40 ± 2.13%; F = 169.00, p < 0.001, η² = 0.80).Sub-Elite athletes, in turn, demonstrated significantly superior performance compared to the Developmental tier in exhaustive load (F = 23.77, p < 0.001, η² = 0.36), number of workload increments (F = 30.77, p < 0.001, η² = 0.42), power at VO₂max (F = 23.32, p < 0.001, η² = 0.36), and GET/VO₂max ratio (F = 55.58, p < 0.001, η² = 0.57), thereby establishing a consistent hierarchical pattern in aerobic performance.

Table 5.

Comparison of physiological and performance variables during incremental rowing testing among competitive tiers

Variables	Elite Tier	Sub-Elite Tier	Developmental Tier
Initial load(W)	256.71 ± 20.02	254.29 ± 23.12	247.71 ± 29.26
Exhaustive load(W)	401.41 ± 23.00**##	367.24 ± 20.82##	328.88 ± 38.19
number of workload increments,	5.8 ± 0.5**##	4.8 ± 0.7##	3.7 ± 0.8
V̇O2max(l·min−1)	5.25 ± 0.52	5.05 ± 0.69	4.66 ± 0.62
V̇O2max(ml·kg−1·min−1)	61.55 ± 6.27	59.88 ± 9.70	56.22 ± 5.30
PV̇O2max (W)	401.41 ± 23.00**##	370.00 ± 24.90##	330.35 ± 33.55
V̇O2max PD (s)	143.24 ± 55.34	104.71 ± 51.37	91.76 ± 42.68
GET (L·min− 1)	4.38 ± 0.38##	3.91 ± 0.58##	3.01 ± 0.70
GET/V̇O2max(%)	85.32 ± 3.93**##	77.36 ± 2.72##	69.40 ± 2.13
V̇Emax(L/min)	193.94 ± 17.56	174.92 ± 18.98	166.88 ± 17.81

*p < 0.05, **p < 0.01 vs. Sub-Elite Tier ; #p < 0.05, ##p < 0.01 vs. Developmental Tier

Table 6.

MANCOVA of physiological and performance variables during incremental rowing testing across competitive tiers

Variables	General F of three tiers	Group effect	Between-group effect
			Elite Tier vs. Sub-Elite Tier F, P, η²	Elite Tier vs. Developmental Tier F, P, η²	Sub-Elite Tier vs. Developmental Tier F, P, η²
Initial load(W)	F = 8.53, P < 0.001, η²=0.72	F = 0.99, P = 0.380, η²= 0.05	F = 1.65, P = 0.618, η²=0.04	F = 1.30, P = 0.782, η²=0.03	F = 0.04, P = 1.000, η²< 0.01
Exhaustive load(W)		F = 35.20, P < 0.001, η²= 0.63	F = 27.42, P < 0.001, η²=0.40	F = 68.84 P < 0.001, η²=0.621	F = 23.77, P < 0.001, η²=0.36
number of workload increments,		F = 40.58, P < 0.001, η²= 0.66	F = 28.22, P < 0.001, η²=0.40	F = 80.37, P < 0.001, η²=0.66	F = 30.77, P < 0.001, η²=0.42
V̇O2max(l·min−1)		F = 2.14, P = 0.130, η²= 0.09	F = 1.58, P = 0.646, η²=0.04	F = 4.22, P = 0.139, η²=0.09	F = 1.53, P = 0.669, η²=0.04
V̇O2max(ml·kg−1·min−1)		F = 1.91, P = 0.161, η²= 0.08	F = 0.85, P = 1.000, η²=0.02	F = 3.81, P = 0.173, η²=0.08	F = 1.97, P = 0.502, η²=0.05
PV̇O2max (W)		F = 31.41, P < 0.001, η²= 0.60	F = 22.33, P < 0.001, η²=0.35	F = 62.00, P < 0.001, η²=0.60	F = 23.32, P < 0.001, η²=0.36
V̇O2max PD (s)		F = 2.86, P = 0.069, η²= 0.12	F = 4.23, P = 0.138, η²=0.09	F = 4.35, P = 0.129, η²=0.09	F = 0.37, P = 1.000, η²=0.21
GET (L·min− 1)		F = 11.33, P < 0.001, η²= 0.35	F = 5.57, P = 0.069, η²=0.12	F = 22.66, P < 0.001, η²=0.35	F = 11.10, P = 0.005, η²=0.21
GET/V̇O2max(%)		F = 94.71, P < 0.001, η²= 0.82	F = 93.45, P < 0.001, η²=0.69	F = 169.00, P < 0.001, η²=0.80	F = 55.58, P < 0.001, η²=0.57
V̇Emax(L/min)		F = 3.59, P = 0.036, η²= 0.15	F = 5.63, P = 0.067, η²=0.12	F = 5.14, P = 0.086, η²=0.11	F = 0.30, P = 1.000, η²< 0.01

Among the covariates, body weight emerged as a significant predictor for initial load, exhaustive load, and power at VO₂max. Height and age also showed significant influences on exhaustive load and power at VO₂max, whereas training years did not significantly contribute to any of the aerobic performance models.

Tier differences in 120-second all-out rowing performance parameters

A MANCOVA, controlling for body weight, height, age, and training years, revealed a significant main effect of competitive tier on performance during the 120-second all-out test (F = 5.53, p < 0.001, η² = 0.39). Subsequent univariate analyses confirmed significant tier effects on peak power (PP), mean power (MP), and stroke rate .

Bonferroni-adjusted post-hoc comparisons detailed these differences (see Tables 7 and 8). Elite athletes generated significantly higher PP (524.35 ± 36.31 W) than both Sub-Elite (475.76 ± 56.00 W; F = 10.50, p = 0.007, η² = 0.20) and Developmental athletes (445.06 ± 53.69 W; F = 15.84, p = 0.001, η² = 0.27). For MP, a complete performance gradient (Elite > Sub-Elite > Developmental) was observed: Elite athletes (440.94 ± 25.89 W) surpassed both Sub-Elite (398.18 ± 38.50 W; F = 16.56, p = 0.001, η² = 0.27) and Developmental athletes (353.24 ± 41.29 W; F = 34.85, p < 0.001, η² = 0.44), and Sub-Elite athletes, in turn, outperformed the Developmental tier (F = 9.73, p = 0.010, η² = 0.18).

Table 7.

Comparison of power output in 120-s all-out rowing test among competitive tiers

Variables	Elite Tier	Sub-Elite Tier	Developmental Tier
PP (W)	524.35 ± 36.31**##	475.76 ± 56.00	445.06 ± 53.69
MP (W)	440.94 ± 25.89**##	398.18 ± 38.50##	353.24 ± 41.29
FI (%)	24.8 ± 6.0	24.1 ± 8.6	31.9 ± 9.9
Stroke rate (spm)	37.0 ± 3.0**##	34.0 ± 2.8	32.6 ± 2.3
BLapeak (mol/L)	13.27 ± 1.98	13.50 ± 2.56	12.73 ± 3.06

*p < 0.05, **p < 0.01 vs. Sub-Elite Tier ; #p < 0.05, ##p < 0.01 vs. Developmental Tier

Table 8.

Results of the MANCOVA for power output measures during the 120-s all-out rowing test across competitive tiers

Variables	General F of three tiers	Group effect	Between-group effect
			Elite Tier vs. Sub-Elite Tier F, P, η²	Elite Tier vs. Developmental Tier F, P, η²	Sub-Elite Tier vs. Developmental Tier F, P, η²
PP (W)	F = 5.53, P < 0.001, η²=0.39	F = 9.01, P < 0.001, η²=0.30	F = 10.50, P = 0.007, η²=0.20	F = 15.84, P = 0.001, η²=0.27	F = 2.86, P = 0.294, η²=0.06
MP (W)		F = 18.32, P < 0.001, η²=0.45	F = 16.56, P = 0.001, η²=0.27	F = 34.85, P < 0.001, η²=0.44	F = 9.73, P = 0.010, η²=0.18
FI (%)		F = 0.95, P = 0.395, η²=0.04	F = 0.32, P = 1.000, η²<0.01	F = 0.67, P = 1.000, η²=0.02	F = 1.85, P = 0.536, η²=0.04
Stroke rate (spm)		F = 8.13, P < 0.001, η²=0.27	F = 10.74, P = 0.006, η²=0.20	F = 13.37, P = 0.002, η²=0.23	F = 1.70, P = 0.597, η²=0.04
BLapeak (mmol/L)		F = 0.78, P = 0.45, η²=0.03	F = 0.06, P = 1.000, η²<0.01	F = 0.90, P = 1.000, η²=0.02	F = 1.55, P = 0.660, η²=0.03

Technical execution also differed, as Elite athletes maintained a higher average stroke rate than both lower tiers (all p ≤ 0.006). In contrast, no significant inter-tier differences were detected for the FI or peak blood lactate concentration (all p > 0.395), indicating similar rates of power decline and metabolic extremes across groups despite divergent power outputs.

Among the covariates, only body weight significantly predicted PP and MP. The results establish a clear hierarchy in sustained high-intensity performance, with elite athletes demonstrating superior power output and cadence maintenance.

Discussion

Tier-specific anaerobic power profiles

The analysis of the 30-second all-out test data revealed significant multivariate differences in anaerobic performance profiles across competitive tiers after controlling for key covariates (F = 3.03, p = 0.002, η² = 0.32). This indicates the presence of distinct, tier-specific physiological signatures that persist independently of differences in body size and training experience.

A detailed examination of individual metrics demonstrated that elite athletes exhibited a clear advantage in key determinants of anaerobic performance. Specifically, the Elite tier produced significantly higher PP than the Developmental tier (p = 0.037, η² = 0.13). This superior PP likely reflects enhanced neuromuscular coordination and rapid force production capabilities, which are critical for generating the explosive propulsion required during the initial sprint phase of a rowing race [19]. Furthermore, the Elite tier also demonstrated significantly higher MP compared to the Developmental tier (p = 0.022, η² = 0.15). This finding aligns with previous research by Cerasola et al. [4], which identified mean power during short all-out tests as a robust indicator of rowing performance, thereby reinforcing MP as a key discriminant of competitive level.

The most pronounced inter-tier disparity was observed in the FI, which exhibited a clear hierarchical progression (Elite < Sub-elite < Developmental, with p ≤ 0.017 for significant pairwise comparisons). The lower FI (i.e., less power decline) in elite athletes signifies a superior ability to maintain power output during a maximal 30-second effort, reflecting a highly adapted tolerance to the metabolic perturbations of extreme glycolytic stress, which may encompass more efficient energy metabolism, superior buffering capacity, and enhanced ionic regulation.

Within this context, peak blood lactate (BLa_peak), measured as an indicator of the overall involvement of the anaerobic glycolytic system, was significantly elevated in the Elite tier compared to both the Sub-elite (p = 0.031, η² = 0.14) and Developmental tiers (p = 0.023, η² = 0.15). This elevated BLapeak suggests that elite athletes can sustain a higher glycolytic flux and potentially recruit a larger muscle mass for high-intensity anaerobic work [20, 21], which is consistent with their ability to sustain a higher mean power and exhibit less power decline (lower FI).

Notably, the absence of significant inter-tier differences in stroke rate suggests that technical execution and pacing strategy during the all-out effort were consistent across groups. This finding strengthens the conclusion that the observed performance differentials are primarily rooted in physiological rather than technical factors. Among the covariates, body weight was a significant predictor of both PP and MP (p < 0.001), confirming its important role in power generation inherent to the sport. However, the persistence of significant tier effects after controlling for body weight underscores the substantial contribution of other physiological adaptations.

In summary, the results of this study delineate a comprehensive anaerobic profile characteristic of elite performers: the ability to generate high instantaneous and sustained power outputs, coupled with a remarkable capacity to maintain power during a short, maximal effort (as indicated by a lower FI). From an applied perspective, these findings suggest that training interventions for developing athletes should extend beyond the sole pursuit of power development. A concerted emphasis on cultivating the physiological adaptations that underpin the tolerance to high glycolytic stress and power maintenance is paramount for bridging the performance gap with the elite tier.

Tier-specific signatures in aerobic metabolism

Traditionally, V̇O₂max has been regarded as a central determinant of endurance performance. However, after controlling for key covariates such as body size, age, and training years, this study found no significant between-tier differences in either absolute or relative V̇O₂max among elite, sub-elite, and developmental rowers. This finding prompts a re-evaluation of the multidimensional construct of aerobic capacity. It is crucial to emphasize that V̇O₂max is strongly influenced by genetics, and its trainability is often limited in already well-trained athletes [22]. When athletes reach a high level, their V̇O₂max likely approaches their individual genetic ceiling, providing essential context for interpreting the group data in this study.

What then determines the subtle differences in competitive level among athletes with similarly high aerobic capacity? The MANCOVA in this study confirmed a significant overall effect of competitive tier on aerobic performance profiles. Subsequent analyses revealed that load tolerance, power output at maximal effort, and fractional utilization of aerobic capacity were more critical discriminating factors.

Firstly, elite athletes demonstrated a large advantage in exhaustive load (η² = 0.63), following a clear hierarchical gradient. This indicates that under progressively increasing aerobic demands, elite athletes can not only generate higher peak power but also sustain performance longer. This exceptional physiological resilience embodies the integration of their comprehensive aerobic capacity, anaerobic tolerance, and mental fortitude.

A more insightful finding comes from the examination of power output at V̇O₂max (PV̇O₂max). We observed that elite athletes achieved a higher PVO₂max. This metric represents the upper limit of aerobic power production and reflects the integrated capacity of the oxygen transport and utilization systems to support intense work. While distinct from classical movement economy (oxygen cost at submaximal loads), a higher PV̇O₂max indicates a greater ability to translate high levels of oxygen uptake into propulsive power on the ergometer.

However, the most powerful discriminator was the ratio of gas exchange threshold to maximal oxygen uptake (GET/V̇O₂max). Elite athletes operated at a substantially higher fraction of their V̇O₂max during stable, high-intensity efforts. This high fractional utilization of V̇O₂max reflects a series of refined metabolic adaptations, including superior muscle mitochondrial density and function, enhanced fatty acid oxidation capacity, and more robust lactate clearance and recycling mechanisms [23, 24]. This highly developed metabolic profile allows elite athletes to maximize the utilization of their aerobic system, delay the dominant engagement of anaerobic glycolysis during high-intensity exercise, and thereby conserve crucial muscle glycogen reserves for the decisive final sprint in a race.

In summary, this study delineates a more sophisticated and practical model of elite aerobic capability than traditional views. The core of this model is that, based on the foundational physiological capacity of an excellent V̇O₂max, the key functional determinants differentiating internal tiers among highly trained athletes consists of exceptional load tolerance (exhaustive load), superior peak aerobic power (PVO₂max), and highly optimized fractional utilization of aerobic capacity (GET/V̇O₂max). From an applied perspective, for developmental and sub-elite athletes, the training focus should shift from pursuing marginal gains in V̇O₂max alone to systematically enhancing the determinants of aerobic power and threshold (e.g., high-intensity interval training) and improving technical proficiency to optimize power transfer. This represents the most effective pathway to narrow the performance gap with the elite tier.

Tier-specific signatures in sustained maximal effort

The 120-second all-out test reveals a complex, multi-dimensional profile of sustained maximal performance that clearly distinguishes elite rowers. The MANCOVA revealed a significant multivariate effect (η² = 0.39), confirming distinct performance profiles across competitive tiers that persist after controlling for anthropometric covariates. The elite advantage in sustained efforts is characterized not by a single physiological marker, but by an integrated synergy of metabolic regulation, power maintenance, and cadence preservation under fatigue.

The most direct differentiators were the sustained power output and cadence preservation. Elite athletes demonstrated significantly higher PP and MP than both Sub-elite and Developmental groups (all p ≤ 0.007), with MP exhibiting a very large effect size (η² = 0.45). This superior power maintenance was coupled with a significantly higher average stroke rate (η² = 0.27) in the Elite tier (p ≤ 0.006). The ability to maintain a higher stroke rate throughout the 120-second test likely reflects a combination of factors, including superior force production and application per stroke, more efficient recovery mechanics, and possibly different pacing strategies or greater strength reserves [25]. This indicates that elite rowers possess an integrated ability to preserve both muscular power and technical cadence under duress, a signature of highly refined neuromuscular coordination and skill stability even in a fatigued state [26]. Research suggests that such superior neuromuscular coordination and the ability to maintain force production during repeated efforts are key differentiators in athletic performance.

The absence of significant between-tier differences in both FI and BLa_peak (all p > 0.395) in this 120-second test presents a compelling contrast to the 30-second test results. This pattern underscores a fundamental shift in the dominant physiological determinants of performance as effort duration extends. In a 30-second test, which primarily challenges the capacity of the immediate and rapid glycolytic energy systems, a higher BLa_peak is indeed a hallmark of superior anaerobic capacity and high glycolytic flux. However, as the effort extends to 120 s, there is a progressive increase in the contribution of aerobic energy metabolism to the total power output. Within this context, a single measurement of BLa_peak should not be interpreted as a direct index of lactate tolerance or as the primary determinant of PP, given the relatively minor contribution of the lactic acid system to instantaneous PP [26]. Instead, in a 120-second effort, BLa_peak can be more accurately viewed as a net outcome of the overall glycolytic activity and its interaction with aerobic and clearance processes.

We propose that the apparent metabolic convergence (similar BLa_peak across tiers) likely stems from a dynamic balance between lactate production and clearance. It is reasonable to hypothesize that elite athletes, by virtue of their consistently higher absolute power output throughout the test, generate a substantial lactate load. The fact that their final BLa_peak was not correspondingly elevated suggests the potential action of superior lactate clearance and utilization mechanisms, although this study lacked direct measures (e.g., lactate kinetics) to confirm this hypothesis.

Consequently, the similar BLapeak across tiers in the 120-second test is not an indicator of similar glycolytic stress. Rather, a plausible interpretation is that elite athletes operating at a higher metabolic steady state, where a high rate of lactate production is effectively matched by a high rate of lactate clearance and utilization. This efficient management of metabolic byproducts may contributes to their increased ability to maintain power output. From an applied perspective, these findings advocate for a training philosophy that moves beyond simply accumulating lactate. For developing athletes, the focus should be on cultivating the physiological infrastructure that supports high lactate clearance and the neuromuscular resilience to maintain power output and technique under fatigue, thereby bridging the gap with the elite performance signature.

Conclusions

This study delineates clear, tier-specific physiological signatures in male rowers. Elite athletes are characterized by an integrated physiological profile: superior anaerobic power and maintenance (higher PP, MP and lower power decline in a 30-second test), exceptional tolerance to sustained maximal efforts (higher power output and cadence preservation over 120 s), and optimized aerobic function (superior exhaustive-load tolerance, peak aerobic power [PV̇O₂max], and fractional utilization of aerobic capacity [GET/V̇O₂max]). For Sub-elite athletes, the primary constraints appear to be lower glycolytic engagement and reduced aerobic power, whereas Developmental athletes require foundational development across all physiological domains. The three-phase testing protocol effectively discriminates between these tiers, providing a quantifiable, physiology-based profile that can inform the move beyond result-based classification toward more individualized training interventions.

Limitations

This study has several limitations. First, its cross-sectional and discriminative design can identify tier-associated physiological signatures but cannot establish causal relationships or validate a predictive framework. Second, the physiological mechanisms discussed (e.g., fiber-type recruitment, lactate clearance) are inferred from performance data and the literature rather than directly measured via invasive techniques such as muscle biopsy. This approach was necessary given the elite athlete cohort, as repeated traumatic testing is not feasible; our protocol of collecting blood lactate samples only twice per stage was designed to balance scientific insight with practical constraints, informed by our extensive experience in the field. Finally, while the sample is substantial for an elite sport study, its homogeneity and size limit the generalizability of findings and the power to detect smaller inter-tier differences.

Authors’ contributions

Conceptualization, Zhigang Gong; Data curation, Hengjin Lai; Formal analysis, Yuxiao Hu; Funding acquisition, Weiyin Huang; Investigation, Yuxiao Hu, Hengjin Lai and Shiming Qiu; Methodology, Zhigang Gong and Pengcheng Guo; Project administration, Pengcheng Guo; Resources, Pengcheng Guo; Supervision, Weiyin Huang; Validation, Zhigang Gong; Writing – original draft, Yuxiao Hu and Hengjin Lai; Writing – review and editing, Zhigang Gong and Weiyin Huang.

Funding

The authors would like to acknowledge the financial support provided by the China Canoeing Association Technology Service Grant Project (CAA201906) and the Research Project Funded by Jiangxi Provincial Sports Bureau (2015016).

Data availability

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The experimental protocol was approved by the Institutional Review Board of Jiangxi Normal University (IRB-JXNU-PEC-2024019), and all procedures complied with the ethical principles outlined in the Declaration of Helsinki. Informed consent was obtained from all individual participants included in the study.

Consent for publication

Participants signed informed consent regarding publishing their data.

Competing interest

The authors declare no competing interests.