Relationship between running performance and weather in elite marathoners competing in the New York City Marathon

It is well known that weather and pacing have an influence on elite marathon performance. However, there is limited knowledge about the effect of weather on running speed in elite marathoners. The aim of the present cross-sectional study was to investigate potential associations between running speed and weather variables in elite runners competing in the ‘New York City Marathon’ between 1999 and 2019. Data from all official female and male finishers with name, sex, age, calendar year, split times at 5 km, 10 km, 15 km, 20 km, 25 km, 30 km, 35 km, 40 km and finish and hourly values for temperature (°Celsius), barometric pressure (hPa), humidity (%) and sunshine duration (min) between 09:00 a.m. and 04:00 p.m. were obtained from official websites. A total of 560,731 marathon runners' records were available for analysis (342,799 men and 217,932 women). Pearson and Spearman correlation analyses were performed between the average running speed and the weather variables (temperature, pressure, humidity and sunshine). Ordinary Least Squares (OLS) regressions were also performed. The runner´s records were classified into four performance groups (all runners, top 100, top 10 and top 3) for comparison. Differences in running speed between the four performance groups were statistically significant (p < 0.05) for both men and women. Pearson (linear) correlation indicated a weak and positive association with humidity in the top 10 (r = 0.16) and top 3 (r = 0.13) performance groups that the running speed of the elite runners was positively correlated with humidity. Regarding sunshine duration, there was a weak and positive correlation with the running speed of the elite groups (r = 0.16 in the top 10 and r = 0.2 in the top 3). Spearman correlation (non-linear) identified a weak but negative correlation coefficient with temperature in all runners’ groups. Also, non-linear positive correlation coefficients with humidity and sunshine can be observed in the Spearman matrixes. A Multivariate Ordinary Least Squares (OLS) regression analysis showed no predictive power of weather factors. For elite runners competing in the ‘New York City Marathon’ between 1999 and 2019, the main findings were that elite runners became faster with increasing humidity and sunshine duration while overall runners became slower with increasing temperature, increasing humidity and sunshine duration. Weather factors affected running speed and results but did not provide a significant predictive influence on performance.