Modeling within-field variability of turfgrass surface properties and athlete performance

Surface properties of turfgrass sports fields exhibit within-field variability. Wearable global positioning system (GPS) athlete performance tracking units offer the possibility to investigate its impact on performance. These data and R code are used as an example to model the relationship between within-field variability and athlete performance; specifically, speed. Data were collected on one home field from two games involving a collegiate club rugby team. Athlete speed data were collected from GPS athlete performance tracking units worn by the participating athletes during the two games. Soil moisture, soil compaction, surface hardness, and turfgrass quality were measured with GPS-equipped sampling devices from the field prior to each game. Using a geographic information system (GIS), the field’s boundary was digitized and divided into 3x3 m grid cells that were each assigned an ID number. Athlete speed and field data were further manipulated in the GIS to calculate average athlete speed and field property variability scores (-3 to 3, where -3 indicates the lowest valued areas in the field and 3 indicates the highest valued areas in the field) in each grid cell both games. The end result were spreadsheets from each game that contain columns with grid cell ID numbers, data point counts and mean speed of each athlete, and variability scores for each surface property for all cells each game. Data point counts indicated athletes’ time spent within a cell. Counts were used to determine weights for calculating the team’s weighted mean speed in each cell both games, where those who spent the most time in a cell had more weight in the cell’s mean speed. The field was also further divided into larger sections to account for areas that may receive varying amounts of play. Linear regressions were conducted to analyze the data. In the models, a 1-unit increase in field property variability score will correspond to an increase (positive coefficient) or decrease (negative coefficient) in team speed. Depending on the game, within-field variability of each measured surface property, as well as a few interactions, did significantly influence team speed. Athletes location on the field also influenced team speed. This is information is useful for understanding the impact of field conditions on athlete performance. It can be used by coaches, trainers, athletes, and field managers to better prepare for, or manage, turfgrass sports fields.草坪运动场的场地表面特性存在场内变异性。可穿戴式全球定位系统（GPS）运动员表现追踪设备为探究该变异性对运动表现的影响提供了可能。本数据集及配套R代码可作为示例，用于建模场内变异性与运动员运动表现——具体为奔跑速度——之间的关联。 数据采集自某大学俱乐部橄榄球队的两场主场赛事。运动员奔跑速度数据由参赛运动员佩戴的GPS运动表现追踪设备采集获得。每场赛事开展前，研究人员使用搭载GPS的采样设备对场地的土壤含水率、土壤紧实度、表面硬度以及草坪质量进行了测量。借助地理信息系统（GIS），研究人员对场地边界进行数字化处理，并将其划分为3×3米的网格单元，每个单元均分配唯一编号。随后在GIS中对运动员速度数据与场地属性数据进行进一步处理，计算两场赛事中每个网格单元内的运动员平均速度，以及各场地属性的变异性评分（评分区间为-3至3，其中-3代表场地内属性值最低的区域，3代表属性值最高的区域）。 最终产出两场赛事对应的电子表格，其中包含以下列字段：网格单元编号、各运动员的数据点计数与平均速度，以及每场赛事中所有网格单元的各场地属性变异性评分。数据点计数代表运动员在对应网格单元内的停留时长，该计数被用于计算两场赛事中各网格单元内球队加权平均速度的权重——即在某单元停留时长越长的运动员，对该单元平均速度的贡献权重越高。此外，研究人员还将场地划分为更大的区域，以覆盖可能受到不同使用强度的场地范围。本研究采用线性回归方法对数据进行分析。 在所构建的模型中，场地属性变异性评分每提升1个单位，将对应球队奔跑速度的提升（系数为正）或下降（系数为负）。不同赛事中，各测量场地属性的场内变异性，以及少量交互项，均对球队奔跑速度存在显著影响。运动员在场地内的位置同样会影响球队奔跑速度。该研究结果有助于理解场地条件对运动员运动表现的影响，可供教练、训练师、运动员以及场地管理人员用于优化草坪运动场的赛前准备与日常管理工作。