Secondary Logo

Journal Logo

Prediction of Planetary Mission Task Performance for Long-Duration Spaceflight


Medicine & Science in Sports & Exercise: August 2019 - Volume 51 - Issue 8 - p 1662–1670
doi: 10.1249/MSS.0000000000001980

Introduction This study aimed to determine values and ranges for key aerobic fitness variables that can individually map the level of success for planetary mission tasks performance for long-duration spaceflight, with the goal to develop a predictor-testing model that can be performed with in-flight equipment.

Methods We studied a group of 45 men and women who completed a series of mission-critical tasks: a surface traverse task and a hill climb task. Participants performed each mission task at a low and moderate intensity designed to elicit specific metabolic responses similar to what is expected for ambulation in lunar and Martian gravities, respectively. Aerobic fitness was characterized via cycling and rowing V˙O2peak, ventilatory threshold (VT), and critical power. Logistic regression and receiver operating characteristic curve analysis were used to determine the cutoff thresholds for each aerobic fitness parameter that accurately predicted task performance.

Results The participants of this study were characterized by a range of cycling V˙O2peak from 15.5 to 54.1 mL·kg−1·min−1. A V˙O2peak optimal cutoff values of X and Y mL·kg−1·min−1 were identified for the low- and moderate-intensity surface traverse tasks, respectively. For the low- and moderate-intensity hill climb test, the optimal V˙O2peak cutoff values were X and Y mL·kg−1·min−1, respectively. VT and critical power also showed high sensitivity and specificity for identifying individuals who could not complete the mission tasks.

Conclusion In summary, we identified aerobic fitness thresholds below which task performance was impaired for both low- and moderate-intensity mission-critical tasks. In particular, cycling V˙O2peak, VT, and rowing CP could each be used to predict task failure.

Department of Kinesiology, Kansas State University, Manhattan, KS

Address for correspondence: Carl J. Ade, Ph.D., Kansas State University, Manhattan, KS 66506; E-mail:

Submitted for publication August 2018.

Accepted for publication February 2019.

Online date: March 15, 2019

The microgravity environment during spaceflight elicits a cascade of physiological adaptations that creates a problematic state of “spaceflight deconditioning” that may limit an astronaut’s ability to perform aerobically demanding tasks upon return to terrestrial surfaces (e.g., an asteroid, the moon, or Mars). These adaptations occur in all physiologic systems along the O2 transport pathway, with the potential for significant reductions in aerobic exercise capacity that is related, in part, to the duration of microgravity exposure (1). Given that the well-defined state of “spaceflight deconditioning” primarily manifests upon return to terrestrial gravity, a plausible scenario exists in which some mission-critical tasks may become physically challenging enough that a decrease in performance and safety may result. As such, an astronaut may be required to maintain a certain level of cardiorespiratory, cardiovascular, and muscular fitness to complete tasks without exhaustion while maintaining his/her overall well-being.

Flight analog studies have demonstrated that ambulation in a pressurized spacesuit on lunar and Martian gravities places high physical strain on the astronaut, with low walking speeds (<3 km·h−1) eliciting metabolic rates of ~17 and ~28 mL·kg−1·min−1, respectively (2), may elicit an additional ~10 mL·kg−1·min−1 increase due to changes in terrain, topography, etc. (3). Further, based on data across 28 Apollo extravehicular activities (EVA), astronauts showed a mean V˙O2 of 10–14 mL·kg−1·min−1 during surface EVA, and at times they would need to sustain ~20 mL·kg−1·min−1 for at least 20 min. Given that the current minimum fitness threshold requirement of the National Aeronautics and Space Administration (NASA) is a peak oxygen uptake (V˙O2peak) of 32.9 mL·kg−1·min−1 (4), a situation may therefore exist, particularly after long periods in microgravity (1,5), in which these terrestrial activities require a relatively high fraction of, or which exceeds, an individual astronaut’s V˙O2peak, resulting in physical exhaustion. As such, it remains unknown if this current cardiorespiratory fitness standard is appropriate for future terrestrial NASA missions. Owing to this lack of a specific cardiorespiratory fitness threshold for these scenarios, the development of a predictive model that incorporates exercise parameters that can be obtained pre- and in-flight becomes desirable. Therefore, it was our primary aim to determine values and ranges for key aerobic fitness variables that can individually map the level of success for key mission-critical tasks. In addition, we aimed to use this information to develop a predictor-testing model that can be performed with in-flight hardware and astronaut time constraints.

Back to Top | Article Outline



Forty-five healthy participants completed the experiments (Table 1) and were proactively recruited to have a range of low to high fitness levels. Because these tasks would be performed after exposure to microgravity that may decrease cardiorespiratory responses to exercise, we recruited individuals with relatively low peak oxygen uptakes (V˙O2peak) (15–20 mL·kg−1·min−1). We have previously estimated that extended periods of microgravity exposure can elicit decreases in V˙O2peak of >30% (1), highlighting the need to include participants with V˙O2peak values commensurate with the extended spaceflight decay in aerobic capacity. All participants were free of known cardiovascular, pulmonary, or metabolic diseases and were nonsmokers. All subjects gave written consent to participate in the study, which was approved by the Institutional Review Board for Research Involving Human Subjects at Kansas State University and conformed to the Declaration of Helsinki. Subjects were instructed to arrive at the laboratory rested, fully hydrated, and having abstained from vigorous activity for 24 h before testing.



Back to Top | Article Outline

Fitness examination

The fitness examination included cycling and rowing ventilatory threshold (VT) and V˙O2peak, and rowing critical power (CP), which were determined individually over a period of 3 d with at least 24 h between laboratory visits. First, participants completed a graded incremental exercise test on a cycle ergometer (Lode, Groningen, The Netherlands) consisting of 3-min stages at 20, 50, 100, and 175 W or 20, 50, 100, 150 W, depending on the participants’ self-reported level of physical activity, followed by 25-W increments every 1 min to volitional exhaustion. This exercise protocol was chosen as it has consistently been used to evaluate V˙O2peak in preflight and in-flight astronauts for NASA shuttle, Skylab, and International Space Station missions (5,6).

On the second day of testing, participants completed a graded incremental rowing test on a commercially available rowing ergometer (Concept2 Inc., Morrisville, VT), which used a protocol identical with the cycling test. Each participant was trained in rowing techniques and familiarized with the incremental rowing protocol on at least one occasion before the test day. Throughout the cycling and rowing incremental exercise tests, metabolic and ventilatory data were continuously recorded via a gas exchange measurement system (True One 2400; Parvo Medics, Sandy, UT), which was calibrated before each testing session according to the manufacturer’s instructions. For each protocol, V˙O2peak was defined as the highest 15-s value achieved during exercise. Maximal effort was confirmed by attainment of at least three criteria: 1) a respiratory exchange ratio >1.1; 2) a heart rate >90% of age-predicted maximum; 3) a plateau of V˙O2 defined as no expected increases (<150 mL·min−1) in V˙O2 from the previous test stage; or 4) a rating of perceived exertion >17 on Borg’s 6–20 scale. The V˙O2 corresponding to the VT was determined as the V˙O2 at which V˙CO2 increased out of proportion with respect to V˙O2 and there was an increase in E/V˙O2 with no increase in E/V˙CO2 (7). Heart rate was continuously recorded with a telemetric heart rate monitor (FT7; Polar Electro Inc., Lake Success, NY).

On the third day of testing, an all-out test, similar to that used for cycling (8) and running (9), was used to determine rowing CP using a previously published protocol (10). All participants completed an initial all-out familiarization test before the test used for data analysis. For the all-out CP test, participants began by warming-up on the rower at 20 W for 3 min followed by a 2-min rest period. The test was initiated from a preparatory power position, and the participant was instructed to provide a maximal all-out rowing effort for 3 min. Participants were instructed to row as quickly as possible throughout the test, although their stroke rate would decline after achieving an initial peak. Strong verbal encouragement was provided throughout the test. To minimize pacing during the test, participants were blinded to the rowing ergometer’s monitor information and the elapsed time. Power output was recorded stroke by stroke via Bluetooth link with the rowing ergometer’s performance monitor. These data were then analyzed on a laboratory computer, in which CP was defined as the average of the final 30 s of the test.

Back to Top | Article Outline

Critical mission task variables

On randomly ordered days after the cycling and rowing tests, participants completed two critical mission tasks; 1) surface traverse and 2) hill climb. The surface traverse test was designed to simulate a crewmember walking to a rover and loading it with needed supplies. Participants traversed a 1500-m track covered with synthetic rocks and pebbles (<3 cm3) to simulate a surface like that expected on Mars. Walking speed was held constant at 0.85 m·s−1. Upon completion of the 1500-m walk, participants transferred thirty 10-kg simulated cargo transfer bags at a distance of 10 m, one at a time. The hill climb test was designed to simulate a crewmember’s need to climb a hill to deploy a line-of-site communications antenna. Subjects walked at speeds between 55 and 99 m·min−1 on a treadmill for 1000 m with a variable incline (0%–8% grade) equivalent to a 40-m rise in elevation. To simulate carrying a component of the antennae to the top of the hill, subjects carried a “Body Bar” (5.5 kg, 1.2 m). Upon reaching the top, the subject made a return trip consisting of a 1000-m distance with a 40-m descent (no bar).

Each participant performed the surface traverse test and the hill climb test twice. Each test was completed at low and moderate intensities to simulate lunar and Martian environments, respectively. In both instances, the low intensity was achieved by pacing the participants to achieve a steady-state metabolic rate similar to what has been reported for lunar ambulation during the Apollo missions and for simulated lunar ambulation in the Mark III Advanced Space Suit Technology Demonstrator EVA Suit (MKIII) (V˙O2 of ~17–20 mL·kg−1·min−1). The moderate intensity was achieved also by pacing the participants and by the addition of 13.6 kg to the torso and 2.3 kg to each ankle and upper arm (total of 22.8 kg). The pace was determined from a combination of using the previously reported metabolic calculations for walking ambulation (11), experience with this type of pacing in our previously published reports (12), and practice runs in a subset of our cohort. Similar to our previous work (12), the actual metabolic rates required to complete the two tasks at each intensity were determined via breath-by-breath pulmonary gas exchange and ventilation measured using a portable system (Oxycon Mobile; CareFusion Corporation, San Diego, CA) in a subset of subjects (three men and two women). The system was calibrated before each test with gases of known concentration. The volume turbine was calibrated using the manufacturer’s automated flow calibration system according to the manufacturer’s instructions. V˙O2 could not be measured in the entire cohort because of unforeseen technical problems and unavailability of the portable metabolic system. Some individuals declined to perform the moderate-intensity test, which is reported in the results section where appropriate. In all instances, a test was considered failed if subjects could not maintain the required pace or declined to continue due to volitional fatigue.

Back to Top | Article Outline

Statistical analysis

Descriptive statistics were calculated for each parameter obtained during the fitness evaluation and variables obtained from the critical mission tasks. To address the primary aim of determining the feasibility of setting a cardiorespiratory fitness standard, univariate logistic regression analyses were performed on each fitness variable for each critical mission task. A Hosmer–Lemeshow test was used to determine whether the number of finishers and nonfinishers were significantly different from those predicted by the model and then to test the overall model fit. The optimal cutoff values (maximizing the sensitivity, i.e., 100%, but with the highest specificity) for each fitness variable were assessed by receiver operating characteristic (ROC) curve analysis because sensitivity for the detection of task failure was considered more important than specificity. The performance of each fitness variable’s ability to predict task failure was quantified by calculating the area under the ROC curve (AUROC). Note that the ideal AUROC would be equal to 1, with a random guess having an AUROC of 0.5.

To address the secondary aim of developing a practical method for predicting critical mission task performance, the results from the logistic regression analyses were used to formulate nomograms for each cardiorespiratory fitness variable via the rms package of R, version 3.5.0 ( Each developed nomogram is based on proportionally converting each regression coefficient in multivariate logistic regression to a 0- to 100-point scale. The effect of the predictor variable with the highest β coefficient was assigned 100 points. From this nomogram, the points are added across each independent variable to derive a total score that is used to estimate the predicted probability of critical mission task failure. The predicted performance of the nomogram was measured by concordance index (C index) (13). A C index of 0.5 indicates a random predictor, whereas 1.0 indicates a perfect predictor. In all instances, statistical significance was set at P ≤ 0.05, and the results are presented as mean ± SE.

Back to Top | Article Outline


Cohort characteristics

Table 1 provides descriptive data for each fitness evaluation variable. The participants of this study were characterized by a range of cycling V˙O2peak from 15.5 to 54.1 mL·kg−1·min−1. Rowing V˙O2peak had a range from 15.9 to 61.9 mL·kg−1·min−1 and rowing CP from 73 to 236 W. Cycling V˙O2peak and rowing V˙O2peak on average were not significantly different (P = 0.40) and were significantly correlated (r = 0.97, P < 0.001).

Back to Top | Article Outline

Critical mission task performance and ROC analysis

The low-intensity surface traverse test estimated a V˙O2 response of 24.0 ± 1.8 mL·kg−1·min−1 versus the 28.7 ± 1.8-mL·kg−1·min−1 V˙O2 response for moderate-intensity surface traverse test. Six of the 45 participants failed to complete low-intensity surface traverse test, whereas 15 of 39 participants failed to complete the moderate-intensity surface traverse test. Logistic regression analysis revealed cycling V˙O2peak, cycling VT, rowing V˙O2peak, and rowing CP as significant independent predictors of both low and moderate surface traverse performance and that each fitness variable was negatively associated with test failure (i.e., odds ratio <1.0; Table 2; Fig. 1). Table 2 provides the fitness level for each variable at which the predicted probability of failure is ≤20%. In all instances, the Hosmer–Lemeshow test indicated that the number of individuals who failed to complete each test was not significantly different from those predicted by the logistic model and that the overall model fits were good. Results of the ROC analysis with optimal cutoff values for each selected fitness variable at each intensity are illustrated in Figure 2 and summarized in Table 3. On the basis of ROC, each fitness variable was a significant predictor of test performance, with no significant differences between each area under the curve. Importantly, all fitness variables consistently showed both high sensitivity and specificity for identifying individuals who achieved exhaustion during the surface traverse test.









The low-intensity hill climb test estimated a V˙O2 response of 20 ± 2.3 mL·kg−1·min−1 versus the 26.0 ± 0.7-mL·kg−1·min−1 V˙O2 response for the moderate-intensity hill climb test. Six of the 45 participants failed to complete the low-intensity test, whereas 10 of 39 participants failed to complete the moderate-intensity hill climb test. Similar to the surface traverse test, logistic regression analysis revealed multiple significant fitness predictors of test performance (Table 2). Figure 1 illustrates the relationship between the key fitness variables and the probability of test failure. Note that like the surface traverse tests, as fitness level decreases the probability of test failure substantially increases. Furthermore, it is important to recognize that in all cases, the moderate-intensity tests required a higher level of fitness for a given probability of test failure. Similar to the surface traverse test, ROC analysis revealed that V˙O2peak measured during cycling exercise showed the highest sensitivity and specificity for identifying individuals who would not finish the surface traverse test at each intensity.

Back to Top | Article Outline

Development of prediction nomograms

Information from both the surface traverse and the hill climb tests was combined to generate a nomogram that used cycling V˙O2peak, cycling VT, or rowing CP as the primary fitness variable (Fig. 3). In all instances, the nomograms had a C index of >0.92, indicating a strong model fit. For example, to estimate the risk of task failure using cycling V˙O2peak (Fig. 3A), a straight line is drawn from the individual’s V˙O2peak value to the “points” axis (e.g., 35 mL·kg−1·min−1 = 50 points). If the task performed is general ambulation and of high intensity, a line is drawn from the “general ambulation” (general ambulation = 5 points) and “high” intensity (high = 17.5 points) to the “points” axis. The points from the cycling V˙O2peak, intensity, and activity type are added together (72.5 points), and a line is drawn from the “total points” axis to the “probability of failure” axis to determine the estimated probability that this individual will fail the test (72.5 points = < 0.2 probability of failure).



Back to Top | Article Outline


The primary objective of the present investigation was to determine the values and ranges for key aerobic fitness variables that can individually map the level of success for planetary mission-critical tasks. The primary novel finding was that maximal and submaximal parameters of aerobic fitness can be used to predict task failure at multiple task intensities that might be experienced, based on the use of current EVA technology, during Martian- and lunar-based exploration tasks. The use of logistic regression and ROC analyses revealed that as the metabolic demands of the tasks increased, the fitness thresholds that predicted task failure/success also increased. In addition, to identify the suggested aerobic fitness thresholds, we aimed to use this information to develop a predictor-testing model that can be performed with pre- and in-flight hardware and astronaut time constraints. Therefore, using the information from the logistic regression analysis, we developed nomograms to predict task failure. From this analysis, cycling V˙O2peak, VT, and rowing CP could all be independently used within a nomogram to predict task failure, further highlighting the ability of maximal and submaximal fitness variables in mapping critical mission task success.

We have previously demonstrated a significant linear relationship between aerobic fitness and time required to complete simulated exploration type tasks (14). Similar to the current investigation, we demonstrated that V˙O2peak was not the only aerobic fitness parameter associated with performance, but that multiple submaximal parameters also provided valuable information. In particular, critical speed, the running equivalent to CP, and to a lesser extent VT were significant predictors of simulated intralanding site–type activities and a simulated 10-km Walkback scenario. In a similar follow-up study, we further demonstrated that individuals unable to complete simulated exploration type tasks at a V˙O2 similar to that predicted for Mars gravity ambulation achieved a higher %V˙O2peak compared with those who finished (12), highlighting that both an individual’s level of fitness and the relative exercise intensity needed to complete exploration type tasks are important considerations when evaluating mission readiness. In the present study, we expanded on this work by demonstrating that as the intensity or metabolic demands of exploration tasks increase, whether it be due to changes in the gravitational environment (1/6 vs 3/8), suit architecture, or mission scenario (nominal vs off-nominal), so does the minimum level of aerobic exercise fitness required to successfully complete the task.

The current NASA standard for evaluating astronaut readiness is V˙O2peak, with a minimum mission readiness of 32.9 mL·kg−1·min−1 (4). Importantly, to date, this aerobic fitness threshold has been sufficient to allow astronauts to complete EVA tasks during Shuttle and ISS missions. This is because the intensity of these tasks has been very low, with peak metabolic rates of ~1.2 L·min−1 achieved (15). Therefore, the 32.9-mL·kg−1·min−1 standard ensures that astronauts performing an ISS EVA are working at ~30%–40% of their peak cycling aerobic capacity, which is critical given that many ISS EVA are several hours in duration. The very low metabolic demands of Shuttle/ISS EVA, however, are dissimilar to that expected for planetary mission tasks. Data from both Apollo EVA and suited lunar analogs suggest that general mission tasks may require a range of V˙O2 of approximately 10–25 mL·kg−1·min−1 (16). Across 28 Apollo EVA, a mean average V˙O2 of 10–14 mL·kg−1·min−1 was achieved, with the Apollo 15 commander working at a V˙O2 of ~20 mL·kg−1·min−1 for over 20 min (16). It is important to note that the metabolic demands of EVA during the Apollo missions were determined from a combination of V˙O2 estimates from the preflight exercising heart rate-to-V˙O2 relationships, oxygen usage computed from oxygen bottle pressure per minute, and from temperature differences from the water flowing into and out of the astronauts liquid cooling garment, all of which may have under or overestimated the exact metabolic demands of lunar EVA.

In the present study, we simulated a crewmember walking 1500 m to a rover and loading it with supplies at low and moderate intensities. These intensities were chosen to simulate the metabolic rates for lunar (1/6-g) and Mars (3/8-g) ambulation and activities associated with intra- and site-to-site EVA transitions. Using the Mark III Advanced Space Suit Technology Demonstrator EVA Suit (MKIII) prototype suit, treadmill walking at very slow speed (<1 m·s−1), which is approximately the speed expected for intra- and site-to-site EVA transitions, can elicit a V˙O2 of ~17–20 and ~28–30 mL·kg−1·min−1 for lunar and Martian gravitational environments, respectively (2). Although the metabolic data for the MKIII are limited to a suit design with a mass of 121 kg, at these low to moderate speeds, Gernhardt et al. (17) have modeled that an increase in suit mass up to 308 kg (250% of MKIII mass) would only elicit an approximate 3.5 mL·kg−1·min−1 increase in V˙O2 for ambulation speeds expected for intra- and site-to-site transitions (18). Similarly, factors associated with suit inertial mass, stability, and pressure on V˙O2 would be minimal under these conditions (17). Therefore, although the MKIII suit is not an exact replica of NASA’s final terrestrial suit design, at slow ambulatory speeds it provides a prototype design that achieves the metabolic costs associated with a wide array of potential suit designs irrespective of its properties. Using this information, the target metabolic rates used in the present study provide a spectrum of metabolic rates that might occur with terrestrial missions. The low-intensity surface traverse and hill climb tests elicited mean V˙O2 of ~20–24 mL·kg−1·min−1, which is similar to previous lunar estimates in the MKIII (~17–20 mL·kg−1·min−1 (2)) and required a minimum V˙O2peak of ~26 mL·kg−1·min−1 for successful task completion. Although this V˙O2peak successfully maps the completion of these low-intensity tasks, allowing V˙O2peak to reach this low level would result in these types of task being completed at ~75%–85% of V˙O2peak, which may cause undue stress to the astronaut and risk premature fatigue before the completion of the mission task(s). Interestingly, at a lunar EVA V˙O2 of 20 mL·kg−1·min−1, many of the Apollo astronauts were working at ~50% of their preflight V˙O2peak (16). These in-flight recordings were further corroborated with findings that suited individuals ambulate in simulated lunar gravity at a self-selected pace that elicits ~51% of their V˙O2peak (2). Thus, our target V˙O2peak of 26.0 mL·kg−1·min−1 only established the “minimum” threshold for low-intensity tasks like those expected for normal lunar mission tasks, and a higher suggested “target” threshold may be required. For example, setting a suggested “target” threshold at 150% of the “minimum” threshold would allow for low-intensity lunar-type activities to be performed closer to 50% V˙O2peak. Thus, the astronaut can be given both “target” and “minimum” aerobic fitness standard.

Suited ambulation in simulated Martian gravity elicits a V˙O2 that is higher than normal walking on Earth, with even the slowest walking speeds reaching 30 mL·kg−1·min−1 (2). In the present study, the moderate-intensity surface traverse and hill climb tests elicited a mean V˙O2 of ~ 26–29 mL·kg−1·min−1, which is similar to previous estimates in the MKIII (~28–30 mL·kg−1·min−1 (2)). Parallel with the increased mission task intensities were increases in the aerobic fitness thresholds required to successfully complete these tests. This is a critical finding in that the aerobic fitness recommendation given to astronauts must be dependent on the anticipated intensity of the tasks they will be required to do. Thus, a single aerobic fitness standard may not be sufficient for future planetary missions and must take into consideration the physical demands of the task(s). It is also evident from the present investigation that one single aerobic fitness standard does not perfectly predict mission task performance. For example, the estimated mean V˙O2 values for the moderate-intensity tasks were fairly similar values (<10% difference), yet the predicted fitness thresholds are not. Using logistic regression and ROC analysis, the goal was to find the best fitting model that described the relationship between the dichotomous pass versus failure outcome and the aerobic capacity. Therefore, the minor differences in metabolic rate for the two tests resulted in different numbers of participants “failing” the tests, resulting in different prediction models as indicated by different odds ratio, 95% confidence intervals, and specificity values. This is an important finding that highlights that while multiple parameters of aerobic capacity significantly predict task performance, no one fitness test completely predicts all of the variability in performance.

Although the fitness thresholds defined in the present investigation can be used to guide the initial preflight fitness requirements of an astronaut, it may be better suited as a pre-EVA fitness threshold that should be met in-flight. We have previously modeled the expected decreases in V˙O2peak as a function of days in microgravity and suggested that extended periods of microgravity exposure can elicit decreases in V˙O2peak of >30% (1). Therefore, if an astronaut’s V˙O2peak is ~5% above the target V˙O2peak threshold of 35 mL·kg−1·min−1 (preflight V˙O2peak = 36.8 mL·kg−1·min−1), an assumed 30% decay after a long-duration flight to a destination would result in an in-flight V˙O2peak of 25.8 mL·kg−1·min−1, which is now ~26% below the target threshold. This information could prove invaluable in making in-flight decisions on which crew member is aerobically capable of performing certain mission tasks. In addition to the need to account for decays in fitness level in-flight, the exercise hardware capabilities must also be taken into consideration. As such, in the present study, we included fitness parameters obtained on a rowing ergometer similar to what might be available in NASA’s Orion capsule. Importantly, we demonstrated that fitness parameters such as peak power output, which does not require an onboard metabolic gas exchange system, provide significant prediction of mission task success, thus further highlighting the potential utility of using these fitness thresholds to guide in-flight decisions before performing certain mission tasks.

Several experimental limitations should be considered when interpreting the findings from the present study. A primary research methodological constraint was the fidelity of the simulations of the mission architecture and environment that was available for tasks analyses. Ideally, high-fidelity mock-ups and suits would have been available for testing under different gravitational loading conditions. However, currently, these are either unavailable for human performance testing or do not yet exist. Therefore, all testing used low-fidelity testing scenarios. Because final standards will be physiological in nature, we believe that our approach will represent critical fitness values that need to be accommodated regardless of final suit/vehicle/habitat designs. A second methodological constraint is the potential differences between our subjects and terrestrial mission-based astronauts. Our subjects wore athletic apparel, not a pressurized space suit, and performed each test in 1g, not a microgravity environment. To circumvent these constraints, each subject was individually paced to elicit an absolute metabolic response similar to that reported for ambulation lunar and Martian gravity while in the MKIII space suit, which is only a prototype (2,3). Therefore, the findings of the present proposal will translate to a range of space suits with varying characteristics across multiple conditions. Finally, V˙O2 was not constantly measured in all subjects during the traverse and hill climb tests. However, the V˙O2 response for each intensity was determined in a subset of individuals, allowing insight into the general metabolic demands of each task intensity. Caution should be used when directly applying the determined thresholds to all types of mission-critical tasks because, as mentioned above, the fidelity of the simulated missions and environment is limited relative to the expected real-world mission conditions that astronauts will experience. However, given the current paucity of data on mission metabolic demand and the general aerobic fitness required to complete these tasks, we believe valuable insights can still be gained from this type of low-fidelity experimentation.

In summary, we identified aerobic fitness thresholds below which task performance was impaired for both low- and moderate-intensity mission-critical tasks. In particular, cycling V˙O2peak, VT, and rowing CP could each be used to predict task failure. The results of this work may be used 1) to inform decisions regarding the readiness of crewmembers to perform physically demanding exploration tasks on terrestrial surfaces, 2) to guide the development of exercise countermeasures hardware for exploration missions, and 3) to inform preflight aerobic conditioning of crewmembers.

The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation and do not constitute endorsement by the American College of Sports Medicine.

This study was supported by a NASA research grant awarded to C. J. Ade and T. J. Barstow.

Back to Top | Article Outline


1. Ade CJ, Broxterman RM, Barstow TJ. VO(2max) and microgravity exposure: convective versus diffusive O(2) transport. Med Sci Sports Exerc. 2015;47(7):1351–61.
2. Norcross JR, Lee LR, Clowers KG, et al. Feasibility of Performing a Suited 10-km Ambulation on the Moon—Final Report of the EVA Walkback Test (EWT). Washington (DC): NASA; 2009.
3. Norcross JR, Stroud LC, Schaffner G, et al. The effects of terrain and navigation on human extravehicular activity walkback performance on the moon. 79th Annual Scientific Meeting of the Aerospace Medical Association. 2008. Available from:
4. NASA. NASA Space Flight Human-system Standard Volume 1, Revision A: Crew Health (NASA-STD-3001). NASA Johnson Space Center Reports. 2014. Available from:
5. Ade CJ, Broxterman RM, Moore AD, Barstow TJ. Decreases in maximal oxygen uptake following long-duration spaceflight: role of convective and diffusive O2 transport mechanisms. J Appl Physiol (1985). 2017;122(4):968–75.
6. Levine BD, Lane LD, Watenpaugh DE, Gaffney FA, Buckey JC, Blomqvist CG. Maximal exercise performance after adaptation to microgravity. J Appl Physiol (1985). 1996;81(2):686–94.
7. Beaver WL, Wasserman K, Whipp BJ. A new method for detecting anaerobic threshold by gas exchange. J Appl Physiol. 1986;60(6):2020–7.
8. Vanhatalo A, Doust JH, Burnley M. Determination of critical power using a 3-min all-out cycling test. Med Sci Sports Exerc. 2007;39(3):548–55.
9. Broxterman RM, Ade CJ, Poole DC, Harms CA, Barstow TJ. A single test for the determination of parameters of the speed-time relationship for running. Respir Physiol Neurobiol. 2013;185(2):380–5.
10. Cheng CF, Yang YS, Lin HM, Lee CL, Wang CY. Determination of critical power in trained rowers using a three-minute all-out rowing test. Eur J Appl Physiol. 2012;112(4):1251–60.
11. American College of Sports Medicine; Riebe D, Ehrman JK, Liguori G, Magal M. ACSM’s Guidelines for Exercise Testing and Prescription. 10th ed. Philadelphia: Wolters Kluwer; 2018, 472 p.
12. Ade CJ, Broxterman RM, Craig JC, et al. Prediction of lunar- and Martian-based intra- and site-to-site task performance. Aerosp Med Hum Perform. 2016;87(4):367–74.
13. Harrell F. Regression Modeling Strategies. 1st ed. New York: Springer-Verlag New York; 2001. pp. 1–522.
14. Ade CJ, Broxterman RM, Craig JC, Schlup SJ, Wilcox SL, Barstow TJ. Relationship between simulated extravehicular activity tasks and measurements of physical performance. Respir Physiol Neurobiol. 2014;203:19–27.
15. Downs M, Norcross JR, English K, et al. ISS crewmember aerobic capacity compared with metabolic cost of simulated exploration extravehicular activity. (abstract). In: Proceedings of the AsMA Annual Scientific Meeting; 2018: Dallas (TX).
16. Johnston RS, Dietlein LF, Berry CA. Biomedical Results of Apollo. Washington: Scientific and Technical Information Office, National Aeronautics and Space Administration : for sale by the Supt. of Docs., U.S. Govt. Print. Off; 1975. viii, 592 p.
17. Gernhardt ML, Jones JA, Scheuring RA, Abercromby AF, Tuxhorn JA, Norcross JR. Risk of Compromised EVA Performance and Crew Health Due to Inadequate EVA Suit Systems. Washington (DC): In, National Aeronautics and Space Administration, Johnson Space Center; 2009.
18. Chappel S, Norcross JR, Abercromby A, Bekdash O, Benson E, Jarvis S. Risk of Injury and Compromised Performance due to EVA Operations. Houston (TX): National Aeronautics and Space Administration; 2017.


© 2019 American College of Sports Medicine