Effect of ski geometry and standing height on kinetic energy: equipment designed to reduce risk of severe traumatic injuries in alpine downhill ski racing.

BACKGROUND: Injuries in downhill (DH) are often related to high speed and, therefore, to high energy and forces which are involved in injury situations. Yet to date, no study has investigated the effect of ski geometry and standing height on kinetic energy (EKIN) in DH. This knowledge would be essential to define appropriate equipment rules that have the potential to protect the athletes' health.
METHODS: During a field experiment on an official World Cup DH course, 2 recently retired world class skiers skied on 5 different pairs of skis varying in width, length and standing height. Course characteristics, terrain and the skiers' centre of mass position were captured by a differential Global Navigational Satellite System-based methodology. EKIN, speed, ski-snow friction force (FF), ground reaction force (FGRF) and ski-snow friction coefficient (CoeffF) were calculated and analysed in dependency of the used skis.
RESULTS: In the steep terrain, longer skis with reduced width and standing height significantly decreased average EKIN by ∼ 3%. Locally, even larger reductions of EKIN were observed (up to 7%). These local decreases in EKIN were mainly explainable by higher FF. Moreover, CoeffF differences seem of greater importance for explaining local FF differences than the differences in FGRF.
CONCLUSIONS: Knowing that increased speed and EKIN likely lead to increased forces in fall/crash situations, the observed equipment-induced reduction in EKIN can be considered a reasonable measure to improve athlete safety, even though the achieved preventative gains are rather small and limited to steep terrain. Published by the BMJ Publishing Group Limited. For permission to use (where not already granted under a licence) please go to http://www.bmj.com/company/products-services/rights-and-licensing/

RCT Entities:   Population Interventions Outcomes

Introduction

Alpine ski racing is known to be a sport with a high risk of sustaining severe injuries.1 2 Injury rates for World Cup (WC) athletes were found to differ among the competition disciplines, particularly when calculated as injuries per 1000 runs: they increased from slalom to giant slalom, super-G and downhill (DH) with the knee as the most frequently injured body part.1 However, if injury risk is normalised with risk exposure time and calculated as the number of injuries per time skiing, the disciplines giant slalom, super-G and DH can be considered to be equally dangerous but for different reasons.3

With respect to the injury causes, a recent study assessing the skier biomechanics in WC alpine skiing found that injuries in super-G and DH are most likely due to high speeds, jumps and higher workloads caused by long competition times.3 High speed is expected to shorten the preparation time necessary for the skier to adapt to jumps and demanding course sections.3 High speed is also expected to increase jump length and air time, resulting in an increased risk of falling.3 4 Furthermore, high speed and, therefore, high kinetic energy (EKIN=1/2×mass×speed2) are likely to increase the forces that occur at the impact in fall or crash situations.3 Consequently, reducing EKIN can be considered a potential prevention tool, particularly in super-G and DH.3

Thus far, it is known that course setting might be an effective preventative measure to control skier speed and EKIN in steep terrain in DH courses.5 6 In addition, equipment-related measures, namely different ski geometries and standing height (ie, distance from ski base to binding plate cover), might potentially reduce EKIN/speed, as was hypothesised by expert stakeholders of the WC ski racing community.7 Yet to date, scientific knowledge on DH is very limited,3–6 8–10 and no field study has assessed equipment-related, preventative measures in super-G and DH.

Therefore, this study aimed to investigate the effect of modifications in ski geometry (ski length, ski width) and standing height of DH skis on speed and EKIN while skiing a WC DH course.

Methods

Measurement protocol and data collection

Two recently retired (10 months) male WC athletes (age: 34.5±4.5 years; height: 184±2 cm; weight: 98.5±1.5 kg) skied several runs on five different pairs of skis varying in width (W), standing height (H) and length (L). For each skier, 4 runs per ski were considered for the data analysis (ie, a total of 40 runs). The test order of the skis was randomised and snow conditions were monitored. The reference ski (SKIREF) was built according to the International Ski Federation (FIS) equipment rules being valid until Winter Season 2011/2012.11 The specifications of all other skis were defined by an expert group consisting of representatives of the Ski Racing Suppliers Association (SRS), FIS Race Directors and researchers who took into consideration the existing scientific knowledge and practical experience (table 1). All prototypes were constructed by one company under the guidance of SRS, strictly adhered to the predefined geometrical variables (table 1) and material composition.

Table 1

Specification of the basic geometric parameters of the DH skis used for the experiments

	SKIREF*	SKIWH	SKILH	SKIWL	SKIWLH
Width (mm)	69	65	69	65	65
Standing height (mm)	50	40	40	50	40
Length (cm)	216	216	220	220	220

*SkiREF represents the original DH racing skis according to the FIS equipment rules valid until Winter Season 2011/2012.8

DH, downhill; FIS, International Ski Federation; SKIREF, reference ski.

The biomechanical field experiment was conducted on the lower part of the FIS WC DH course in Åre (Sweden), directly after a women's WC DH race. The first section of the course was steep and turning (SectionSTEEP), the second section was flat and less turning (SectionFLAT; figure 1). The analysis for SectionSTEEP started at the first gate where skiers reached 19.9 m/s and ended at gate number 9. The analysis of SectionFLAT started at gate 11 and ended at gate 21.

Figure 1

Map of the course with gates, gate numbers and skier trajectories. The boundaries of the steep section (SectionSTEEP) and the flat section (SectionFLAT) and the sharp right turn at gate 2 are indicated with arrows.

Course setting and the snow surface geomorphology were captured using static differential global navigation satellite systems (dGNSS) and were reconstructed in a digital terrain model (DTM), as conducted in earlier studies.5 6 Each skier's instantaneous three-dimensional position was captured by kinematic dGNSS (50 Hz), using GPS and the Russian (GLONASS) global navigation satellite systems, L1 and L2 signals, and was carried in a small backpack as described in detail in previous studies.12 13 The centre of mass (CoM) position of the skier was approximated using a virtual pendulum model, which was attached to the skier's antenna position and the intersection of the pendulum with the snow surface DTM.12

Postprocessing and parameter calculations

Course setting was characterised by gate distance and horizontal gate distance,14 15 using the definition of double gate turns in speed disciplines introduced in earlier studies.5 6 16 Skier speed, turn radius and EKIN were derived from the CoM position (measurement-system accuracy: 0.1 m).12 Ground reaction force (FGRF) and ski–snow friction force (FF) were calculated by the application of a kinetic model on the CoM position, the virtual pendulum model and the body extension (measurement-system accuracy: 63N for FGRF; 42N for FF).17 The ski–snow friction coefficient (CoeffF) was calculated as the coefficient of FF and FGRF.16 To compare the time series data of speed, turn radius, EKIN, FGRF, FF and CoeffF, between runs, parameters were spatially normalised based on an alternative approach specifically dedicated to the characteristics of competitive alpine DH skiing (see online supplementary data).

Statistical analysis

The statistical analysis was conducted for skiers A and B and for SectionSTEEP and SectionFLAT separately. For each run of each skier and each ski, EKIN averages of the SectionSTEEP and SectionFLAT were calculated. Based on these single EKIN section averages, participant mean±SD values were computed for all tested skis. In addition, EKIN section averages were tested for significant differences between the tested skis using a one-way analysis of variance (ANOVA; p<0.05). For pairwise comparison, the post hoc Tukey-Kramer correction was used. To assess if the response to the ski intervention was similar for both skiers, the average EKIN difference between SKIREF and each ski type was compared between the skiers for each section. To test whether these ski response differences between the skiers were statistically significant, two-sided Student t tests (p<0.05) were used.

Finally, a local subsection in which the equipment-induced effects on EKIN seemed to be the greatest was defined. Within this subsection (the sharp right turn at gate 2), the relation between EKIN and the EKIN explanatory parameters were investigated by the use of: (1) Spearman's rank correlation coefficients between the WLHSKIWLH−SKIREF differences in speed and the WLHSKIWLH−SKIREF differences in FF; and (2) a multiple regression analysis assessing the contribution of differences in CoeffF and in FGRF to explain the local differences in FF between SKIREF and SKIWLH.

Results

General characteristics of the test setup

Table 2 presents the course characteristics of the test setup on the WC DH course. Table 3 shows the average speed and EKIN within SectionSTEEP and SectionFLAT for both skiers. Figure 2 illustrates the instantaneous EKIN (ie, its mean±SE from start of SectionSTEEP to end of SectionFLAT for both skiers using SKIREF). It can be visually observed that EKIN/speed was higher in SectionFLAT than SectionSTEEP, and that instantaneous differences between the skiers were present at various locations of the course.

Table 2

Characteristics of the course for the steep (SectionSTEEP) and the flat (SectionFLAT) sections of the downhill course

Parameter	Entire course
Course length (m)	1302
Vertical drop (m)	402
Number of gates ( )	21
Mean run time (s)	50.0
	SectionSTEEP	SectionFLAT
Median terrain inclination (°)	−23	−15
Median gate distance (m)	84.23	61.67
Median horizontal gate distance (m)	35.65	12.23
Mean direction change from gate to gate (°)	23	11

Table 3

Characteristics of speed and kinetic energy (EKIN) for the course sections steep (SectionSTEEP) and flat (SectionFLAT)

Parameter	SectionSTEEP		SectionFLAT
Group mean of EKIN (J/BW)	30.9		44.7
Group mean of speed (m/s)	24.6		29.6
	Skier A	Skier B	Skier A	Skier B
EKIN (J/BW)	31.1	30.6	45.3	41.1
Speed (m/s)	24.7	24.5	29.8	29.4

Figure 2

Areas around the estimate of the mean (±SE) illustrating instantaneous kinetic energy (EKIN) for both skiers skiing on reference ski. Grey lines: skier A; black lines: skier B. The boundaries of the steep section (SectionSTEEP), the transition between sections and the flat section (SectionFLAT) are indicated at the top of the figure.

Differences in EKIN between the tested skis within specific sections of the course

The results reporting the EKIN section average differences between the tested skis are shown in table 4. The one-way ANOVA analysis was significant for both skiers for SectionSTEEP, but not for SectionFLAT, indicating that the equipment intervention did not have an effect on EKIN/speed in SectionFLAT. At the pairwise comparisons (right side of table 4), a difference is negative if the modified skis showed smaller EKIN mean values than SKIREF. The only ski prototypes that caused a statistically significant reduction in EKIN/speed compared with SKIREF was SKIWLH. This finding was independently observed for both skiers in SectionSTEEP only (skier A: −0.95 J/BW, −3.0%; skier B: −1.0 J/BW, −3.2%).

Table 4

Top: statistical analysis comparing the EKIN for skiers A and B in the SectionSTEEP. Bottom: statistical analysis comparing the EKIN for skiers A and B in the SectionFLAT

Ski	EKIN (J/BW)					ANOVA	Pairwise comparisons (%)
	SKIREF	SKIWH	SKILH	SKIWL	SKIWLH	p Value	SKIWH−SKIREF	SKILH−SKIREF	SKIWL−SKIREF	SKIWLH−SKIREF
SectionSTEEP
Skier A	31.35±0.30	31.15±0.04	31.44±0.24	31.65±0.37	30.40±0.43	.001***				−3.0*
Skier B	31.08±0.18	31.15±0.04	31.05±0.25	31.2±0.42	30.08±0.55	.005***				−3.2*
SectionFLAT
Skier A	45.24±0.47	45.47±0.61	45.67±1.01	46.32±0.08	45.10±0.64	0.209
Skier B	43.30±0.96	45.07±0.61	44.54±0.15	44.96±1.30	44.70±0.73	0.108

Level of significance: *p<0.05, **p<0.01, ***p<0.001. Post hoc method with Tukey-Kramer correction for pairwise comparison.

The mean±SD of EKIN is given for each prototype and course section on the left side. The right side presents selected pairwise ANOVA comparisons of SKIWH, SKILH, SKIWL and SKIWLH with SKIREF. A difference is negative if the modified skis showed smaller EKIN mean values than when skiing on SKIREF.

ANOVA, analysis of variance; EKIN, kinetic energy; SectionFLAT, flat course section; SectionSTEEP, steep course section; SKIREF, reference ski.

The extent to which a certain pair of skis caused the same EKIN/speed difference with respect to SKIREF was not significantly different between skiers but was significantly smaller in SectionSTEEP (0.07±0.03 m/s) than in SectionFLAT (0.21±0.08 m/s).

Differences in EKIN between SKIWLH and SKIREF in SectionSTEEP

Figure 3 illustrates the instantaneous differences in EKIN between SKIREF and SKIWLH for skiers A and B within SectionSTEEP. The specific subsection ‘sharp right turn’ starts at the first gate, where mean speed is equal for both skiers and skis (skier A: 19.9 m/s with SKIREF and 19.9 m/s with SKIWLH; skier B: 20.0 m/s with SKIREF and 20.0 m/s with SKIWLH), and ends when the average turn radius of both skiers and skis exceeded 125 m after the passage of gate 2. Within this subsection, the percentage differences in EKIN between SKIWLH and SKIREF were −3.6% for skier A and −7.0% for skier B.

Figure 3

Top (skier A) and middle (skier B): areas of uncertainty around the estimate of the mean (±SE) illustrating instantaneous kinetic energy (EKIN) and reference ski (SKIREF) versus SKIWLH within steep section. Black area: SKIREF; grey area: SKIWLH. Bottom: instantaneous differences between SKIREF and SKIWLH for skier A (black line) and for skier B (grey line). The dashed lines indicate the sharp right turn at gate 2.

Table 5 shows the Spearman's rank correlation coefficients describing the relation between the SKIWLH−SKIREF differences in speed and the corresponding differences in FF during the sharp right turn in the steep section for skiers A and B. For skier B, speed was significantly correlated with FF.

Table 5

Spearman's rank correlation coefficients describing the relation between the SKIWLH−SKIREF differences in speed, and the differences in FF during the sharp right turn at gate 2 for skiers A and B

Skier A	Speed
FF	0.800n.s.
Skier B	Speed
FF	1.000**

Level of significance: nsnot significant at p<0.05, *p<0.05, **p<0.01, ***p<0.001.

FF, ski–snow friction force; SKIREF, reference ski.

Table 6 shows the results of the multiple regression analysis assessing the contribution of the WLHSKIWLH−SKIREF differences in CoeffF and in FGRF to explain the WLHSKIWLH−SKIREF difference in FF during the exemplary sharp right turn. This analysis indicated that the differences in CoeffF were more important than those in FGRF to explain the differences in FF (figure 4).

Table 6

Results of the multiple regression analysis assessing the contribution of differences in the average ski–snow friction coefficient (CoeffF) and in the average ground reaction force (FGRF) to explain the difference in average ski–snow friction force (FF) between reference ski (SKIREF) and SKIWLH during the sharp right turn at gate 2

Predictors of FF	β-weight	p Value
Skier A
CoeffF	1.011	0.004
FGRF	0.123	0.011
Skier B
CoeffF	0.944	0.006
FGRF	−0.320	0.019

Model for skier A: adjusted R2=1.000; p=0.012.

Model for skier B: adjusted R2=1.000; p=0.010.

Figure 4

Areas of uncertainty around the estimate of the mean (±SE) illustrating speed, ski–snow friction force (FF), ground reaction force (FGRF) and ski–snow friction coefficient (CoeffF). Black area: reference ski; grey area: SKIWLH. The in-depth analysis of speed and FF, as well as FGRF and CoeffF for skier A (left side) and skier B (right side) are presented for the exemplary sharp right turn in steep section. G1 and G2 indicate gates 1 and 2, TS and TE indicate turn start (when mean turn radius across both ski types falls below 125 m) and turn end (when mean turn radius across both ski types exceeds 125 m).

Discussion

The main findings of the study were that no difference between the tested skis on average EKIN in SectionFLAT were observed, but SKIWLH caused a ∼3% EKIN reduction for both skiers in SectionSTEEP compared with SKIREF. No other differences between SKIREF and the ski prototypes were observed. The largest reduction in EKIN was found in a sharp right turn at gate 2 for skier B (−7.0% for skier B; −3.6% for skier A). For skier B, this reduction in EKIN/speed can be explained by increased FF, which was mainly a result of increased CoeffF.

The progress of EKIN over the entire DH course

The current study revealed that for both skiers EKIN/speed was lower in SectionSTEEP than in SectionFLAT (table 3), which is in line with previous findings in men's WC alpine skiing.5 6 18 Moreover, based on the study findings presented in tables 2 and 3, speed, terrain and course setting can be considered representative for both female and male WC races.5 6

Comparing the individual progressions of EKIN/speed between skiers A and B when skiing on SKIREF, it seems that the general characteristics correspond well, but locally differ at certain spots along the course. A similar observation was already reported for other energy-related parameters in giant slalom by Supej.19 Additional analysis revealed that these interindividual differences in EKIN/speed on SKIREF were generally larger than the differences caused by the ski intervention. The agreement in response to the ski intervention between skiers was best in SectionSTEEP where the significant speed reductions were found between SKIWLH and SKIREF for both skiers. These two aspects might strengthen faith in the findings of this study.

The effect of ski geometry and standing height on EKIN

Effects over the entire DH course

As shown in this study, none of the tested pairs of skis caused significant differences compared with SKIREF in SectionFLAT. However, a significant reduction of ∼3% in EKIN was found for the prototype with all three parameters altered (SKIWLH) in SectionSTEEP (table 4). Furthermore, postanalysis for SectionSTEEP revealed that the tested ski interventions had the smallest effect on EKIN/speed in the traverse (ie, between gates 4 and 5). Hence, the combination of steep terrain and turning courses seems to provoke differences between the modified skis and SKIREF, while flat terrain in combination with ‘gliding turns’ (ie, turns, which can be skied in a tucked position) do not.

Local effects within the steep section

Within SectionSTEEP SKIWLH caused an average reduction in EKIN of ∼3% for both skiers. Locally, this reduction was sometimes even larger. A maximal reduction in EKIN (∼−7.0%) was found during the sharp right turn at gate 2 for skier B. Within this specific subsection, the terrain was the steepest and turn radius was the smallest across the entire course. These findings illustrate the extent of the equipment-induced EKIN reduction can—locally—reach quite substantial magnitudes that are comparable to the ones achieved by course setting interventions.5 6 14

For skier B, the local reduction in EKIN during the sharp right turn at gate 2 was explained with a significant correlation between speed reduction and FF, indicating that the loss in speed was mainly a result of higher ski–snow friction. An additional postanalysis from gate 1 to turn end after gate 2 found no difference in the CoM trajectories and turn radii as long as the skiers were turning (turn radius smaller than 125 m), neither between SKIWLH and SKIREF within the same skier, nor between the skiers. Hence, it is reasonable that the differences in the skier's response to the equipment intervention at gate 2 are most likely not a result of different trajectories, but rather due to differences in the ski–snow interaction.

Explanation of the observed local effects in FF

The analysis of how the observed effects in FF can be explained by variables related to the ski–snow interaction revealed that the differences in CoeffF contributed to a greater extent to the differences in FF, than FGRF. Hence, even the most extreme ski prototype (SKIWLH) might not change the skier's movement patterns such that FGRF is substantially affected, but changes how the skis interact with the snow. For skier B the increase in CoeffF and, therefore, in FF seem to start early during the initiation phase of the turn (figure 4), and might be explained by an increased skidding prior to gate passage, as already observed for slalom skiing.20 For skier A the increase in CoeffF was delayed and had a different pattern (figure 4).

Ski geometry and standing height in the context of injury prevention

High speed and, therefore, high EKIN have several aspects that might influence the risk of severe traumatic injuries in super-G and DH.

First, in fall or crash situations, speed is a crucial factor, since the energy that is dissipated by forces during the impact increases with speed by the power of 2. In this context, the observed equipment-induced reduction in EKIN/speed would theoretically lead to the same per cent wise reduction in the impulse (force over time) that acts in impact situations. Restated, a 3–7% local reduction in EKIN would result in a reduction of the impact forces by 3–7% if the impact process time is held constant. However, it has to be pointed out that such a preventative gain can only be achieved when skiing on the most extreme ski prototype (ie, varying in width, standing height and length) and probably only within steep terrain, particularly if course setting causes small turn radii.

Second, anticipation and adaptation time within demanding course sections decrease with increasing EKIN/speed, which plausibly results in a higher risk for technical and tactical mistakes.3 Thus, a reduction in speed would give athletes more time to prepare for difficult course sections (eg, jumps, rough terrain transitions or turns) and make appropriate technical and/or tactical decisions. However, the observed reduction in speed in this study would only lead to marginal changes in preparation and adaption time. Given that an athlete skis with the average speed of SectionSTEEP (26.4 m/s) and oversees the upcoming course 20 m at a time, the observed equipment-induced reduction in speed would increase adaptation time by only 1/100–3/100 s. Hence, the preventative gain regarding this aspect is limited.

Third, increased EKIN/speed is known to increase jump distance and airtime and might result in severe consequences in case the skier makes mistakes at the take-off.3 4 21 22 Concerning this aspect, the current study findings indicate that only if a jump is located directly within or after a steep section skis modified in length, width and standing height would markedly reduce EKIN/speed. In these situations, a ∼3% reduction in EKIN/speed would reduce jump distance and air time by ∼0.5 m and ∼2/100 s, depending on the shape of the jump.3 4 Since many severe injuries are known to occur at jumps,23 this reduction—even though it is small—might serve as a certain preventative gain to protect the athlete's health.

Compared with the magnitude of the EKIN reduction, which might be achieved by course-related measures,5 6 14 the equipment-induced 3% reduction in EKIN in steep terrain and, therefore, the changes in preparation and adaption time to course features and the changes in jump distance and airtime seem to be rather small. However, by a smart combination with course-related and terrain-related measures, these small effects might become more relevant. Future studies should assess combined preventative measures including equipment and course setting modifications and investigate their effect on EKIN/speed. Knowing that only the most extreme equipment modification showed a preventative effect with respect to EKIN, and considering that the athletes reported a delayed reaction in general and reduced rebound in the second part of the turn for SKIWLH and SKIWH, ski ability and external attractiveness of the sport have to be considered. Benefits and costs must be balanced from an implementation point of view.

Methodological considerations

Since the study suffers from the small sample size and the limited number of repetitions per ski, the findings need to be interpreted with caution. Nevertheless, at this early stage of knowledge on injury prevention measures in DH and super-G, the new technology applied to top-level athletes in the current study might provide insights in the general mechanisms of ski–snow interaction and the effect of modified equipment in the context of injury prevention. Hence, this study might serve as guidance for future studies in a field where previously no scientific knowledge was available.

Conclusion

Recent studies suggested that measures to prevent injuries in super-G and DH should aim at reducing speed and EKIN at spots where skiers are likely to crash. This is the first scientific study assessing equipment-related measures to reduce speed and EKIN in DH. The study revealed that a simultaneous decrease in ski width and standing height and an increase in ski length reduces EKIN by 3%, but within steep terrain only. Locally, even larger EKIN reductions of up to 7% were observed. This indicates that an equipment-induced reduction of EKIN/speed is feasible and can be considered an efficient way to increase the athlete's safety in steep terrain. However, it must be stated that the preventative gain from modified DH skis is limited compared with other external preventative measures, such as course setting. Therefore, effective prevention strategies in DH should involve several different preventative measures that aim to radically slow down skiers at locations where crashes are likely to occur.

This study investigates for the first time the influence of ski equipment-related preventative measures (ski, width, standing height, ski length) on the risk factors kinetic energy and speed in downhill (DH).

It adds more detailed insights into how ski geometry and standing height influence kinetic energy within an entire DH course.

In steep terrain, longer skis with decreased width and standing height reduce average kinetic energy by ∼3%. Locally, the magnitudes of energy reduction were even larger (up to ∼7%).

Speed and kinetic energy of the skier determine to a large extent the impact forces during fall and crash situations. Therefore, the observed equipment-induced reduction in kinetic energy can be considered an effective way to improve the athletes’ safety, although compared with other external interventions, the preventative gain of modified DH skis is rather small.