游乐场树木对儿童游戏行为及健康的影响

著：（格鲁尼亚）玛丽亚米·马格莱克利兹 姜珊 宋阳 黎东莹 译：李晓捷

1 研究背景

1.1 接触大自然对儿童发展的益处

现代社会中儿童的户外游戏和与自然的接触急剧减少，这不仅剥夺了他们游戏的权利，而且会导致“自然缺失症”。这一概念描述了儿童感官能力的退化、注意力集中困难，以及身体和情感疾病的高发率[1]。越来越多的科学研究揭示了儿童日益增长的室内生活方式对健康产生的负面影响，例如儿童肥胖率的快速上升和几种主要非传染性疾病风险的提高[2-3]①。《亲生命假说》（The Biophilia Hypothesis）一书从进化的角度强调了与自然接触对健康的好处；人类有着与生俱来的与自然和其他生物体相互关联的偏好[4]。Kahn和Kellert以亲生命性的观点，在他们的著作《儿童与自然：心理学、社会文化和进化研究》（Children and Nature: Psychological, Sociocultural and Evolutionary Investigations）中探讨了儿童在自然中玩耍的心理、智力和社会文化价值[5]。接触大自然对于培养儿童的整体健康状况和促进其早期发展至关重要。事实上，已有研究证据表明年轻一代接触大自然对健康有广泛的好处，如减轻压力、预防视力问题，以及缓解多动症（注意缺陷多动障碍）患儿的症状[6-9]。多篇文献综述和相关书籍论证了自然接触对儿童的益处[10-13]。

1.2 游戏和游戏行为

游戏是儿童发展过程中的重要组成部分。经典的游戏理论由4个分支组成，从根本上阐释了儿童游戏行为的原因和功能，包括：1）剩余精力说[14]；2）松弛说[15]；3）生活预备说[16-17]；4）复演理论[13,18]。现代游戏理论认为游戏能促进认知和抽象能力，它为游戏在儿童发展中的关键作用奠定了理论基础，并且通过实证研究支持了这些理论与基本概念，衍生出心理分析、警觉调节理论、元交际和认知理论等分支[13-19]。幼儿利用游戏作为学习周围世界的主要来源，交流他们的情感和感受，发展认知思维和推理技能，并练习社交和心智运动的技能[20]。玩耍也涉及个性的发展，它有助于处理人际关系，激发创造力，增加生活的乐趣，并促进学习[21-22]。

游戏有很多种形式。当孩子们在成年人设定的环境（比如游戏规则或游戏设备）中玩耍，他们可能会经历一种有引导的游戏[23]；相对应的，自由玩耍是一种无组织的、自愿的、由孩子发起的活动，允许孩子发展他们的想象力，同时探索和体验他们周围的世界[24]。儿童的游戏行为因年龄、性别和儿童发育阶段等因素而异。根据Piaget对儿童认知发展的3个连续阶段的分类，游戏行为可以根据游戏仅是纯粹的感知运动，抑或与思维本身有一定关系的程度来分类：1）练习性游戏，不断进行重复肌肉运动，不管有没有具体的物体；2）建构性游戏，包括利用物体来建构或创造某些东西；3）戏剧性游戏，儿童用想象的、戏剧化的情景来满足他们的个人愿望和需要；4）规则性游戏，要求儿童能够理解预先设定的规则和对这些规则的调整[25-27]。在社会交往方面，Parten发现学龄前儿童的社会参与程度随着年龄的增长按以下顺序发展：无目的行为、独自玩耍、旁观行为、平行游戏、联合游戏以及合作游戏[28]。Rubin开发了游戏观察量表（Play Observation Scale, POS）来描述这2个游戏层级（即社交游戏行为和认知游戏行为）[26]。POS已经成为许多游戏研究的理论基础和数据收集工具[29-30]。

1.3 游乐场内的自然环境及树木

基于游戏理论和亲近大自然对健康的显著好处，将自然元素融入游乐场设计已经成为促进儿童早期发展的一种有效的干预手段。游乐场内树木的益处包括：1）对儿童早期的认知和智力发展有显著的贡献[31]；2）提高儿童的身体活动水平，增加户外游戏时间[32-34]；3）鼓励幼儿在游戏中的社会互动和情感发展[31,35-36]。游乐场上的自然元素与该场所的疗愈效能具有直接相关性[37]。孩子们喜欢的户外游乐场的自然元素包括：1）水；2）植被；3）动物；4）土壤材料，例如沙子和泥土；5）自然色彩、质地、多样性和变化；6）有遮阴的座位；7）能够提供隐私和视野的不同层次的小空间和空间缝隙；8）可以根据他们的想象力改变的结构、设备和材料，其中包括大量能用来玩耍的零零碎碎（loose parts）[38]。

树木，是景观设计的基本元素，在幼儿的游乐场中可以发挥重要作用。例如，树荫可以通过降低表面温度和减少紫外线辐射提高游乐场的安全性，并提高儿童在玩耍时的整体户外热舒适性[39-40]。根据美国皮肤病学会的研究，1/5的美国人一生中可能会患上某种形式的皮肤癌，5次及以上的晒伤会使患皮肤癌的风险增加一倍。利用树荫，以及设计游戏设施作为提供阴凉的手段（如架高的结构或树荫），并通过人为构建创造更多的阴凉可以帮助保护儿童的皮肤免受阳光照射[41]。在近期的一项研究中，Anderson和他的同事调研了位于澳大利亚悉尼市的139个城市游乐场，并进行了1 033次树荫状况观察。他们发现，在社会经济条件较低地区的游乐场，主要游戏区域的遮阴覆盖率明显比社会经济条件较高地区的平均遮阴覆盖率低34%左右，这引起了学界对预防皮肤癌的关注，并提示须将社会经济层面的因素纳入游乐场树木情况的评估研究当中[42]。

2 研究问题与方法

大量的研究已经强调了树木和树荫在改善游乐场的热舒适性、阳光直射和小气候条件方面的好处。然而，对于在城市游乐场中设置树荫的行为效益和社会效益的探索还比较欠缺。文中记录了2项最新的科研成果，均为城市游乐场的树木和树荫及其对儿童游戏行为影响的案例研究。案例研究Ⅰ比较了在美国华盛顿特区市中心的2个游乐场上，树木覆盖和树荫是如何影响幼儿的游戏行为的。案例研究Ⅱ调查了美国俄克拉荷马州塔尔萨市中心的一个主题儿童公园内的2个游乐场，树木覆盖、树荫和其他自然要素是如何影响幼儿的户外游戏时间、游戏行为和活动模式的。

2项研究均以直接观察为主要方法。根据Rubin开发的游戏观察量表对幼儿的游戏行为进行观察和绘制[26]。案例研究Ⅰ借助了地理信息系统（GIS）云软件来追踪和绘制游戏行为图。其中的其他变量包括2个游乐场从激光雷达数据和EnergyPlus软件的天气数据中获得的每小时日照强度，以及被观察者与自然的视觉距离。案例研究Ⅱ采用传统的纸笔技术对儿童游戏行为进行现场实时跟踪，并记录了儿童在游乐场上的停留时间和日照情况。

1 游乐场 A 平面图Playground A site plan

2 游乐场B平面图Playground B site plan

3 地理信息系统云软件作为研究I中的数据收集工具GIS Cloud as a data collection tool in Study I

3 案例研究Ⅰ

3.1 研究场地及特征

位于美国华盛顿特区市中心的2个户外游乐场被选为案例研究Ⅰ的场地（图1～2）。游乐场A和B均位于城市的东北部，相距约4.8 km（3英里）。这2个游乐场都建于2013年，空间设计和游乐器械设计都是由美国明尼苏达州同一个游乐场设计公司设计的。这2个游乐场的设计概念融合了数学、艺术、建筑、科学和自然。这2个游乐场都提供能容纳2个年龄组的游戏设施：幼儿/学龄前儿童（2～5岁）和儿童（6～12岁）。场地的游戏设备包括跷跷板、感官游戏中心、小屋、桥类器械、攀爬架和秋千。游戏景观设计中的色彩运用受到了大自然的启发，包括丹宁色、湖蓝色、叶绿色、柠檬色、松绿色和棕褐色。游乐场A内有大型遮阴乔木，为游乐场的主要游乐区提供天然遮阴。孩子们可以从主要的游乐区直接接触到几块小草坪和低矮的灌木丛。相比之下，虽然游乐场B的周边种植了一些树木，但主要的游戏区域并没有任何树冠遮阴。

3.2 仪器和观察规程

数据收集是在POS[26]的指导下完成的。POS规定观察者对每一个目标儿童观察10 s（即一个观察区间），然后在接下来的5～10 s内对该儿童的主要游戏行为进行解读并录入记录表。因此，需要1.5～2 min才能获得1 min的观测记录。为了对儿童的游戏行为进行有效的测量，POS建议在任何一天对同一个目标儿童最多只记录共5 min的行为。POS还建议收集至少15 min的POS数据用于分析和结果阐释。因此，在此案例研究中，每个目标儿童需要被观测6个观察区间。每轮连续的观测和记录需持续约30 min，即为一个观察期。根据POS的规定，在每个10 s的观察区间内，只有一个游戏行为应被记录。如果在10 s的观察区间内发生了多个行为，则对占用区间内大多数时长的游戏行为进行记录。如果若干游戏行为的时间长度相同，观察者应该遵循一个递进的原则来进行录入：任何群体的游戏行为高于所有其他行为（例如，群体游戏>平行游戏>独自游戏）。对于认知游戏行为来说，有规则的游戏凌驾于其他结构性或练习性的游戏行为之上（例如，规则性游戏>戏剧性的游戏>探索性游戏>建构性的游戏>功能性游戏）[26]。

除了POS中描述的社交和认知游戏行为外，儿童是否在游乐设施上玩耍也被记录下来。在本研究中，任何在游乐设施上进行的游戏行为都被定义为“有引导的游戏”，而那些在设备区以外进行自我导向游戏的行为则被定义为进行“自由游戏”。

GIS Cloud是一个基于GIS的网络实时地图工具。观察者能够在任何观察区间内在该软件中标注目标儿童的位置，并在其中嵌入的游戏行为记录表上记录儿童的主要游戏行为。图3描绘了GIS Cloud以及行为记录表的用户界面。

3.3 其他变量

除POS中规定的常规变量外，本研究还测量了另外2个变量，即日照/阴影强度以及儿童与游乐场上自然元素的视觉距离。

1）日照强度，指观察期间在游乐场上任何位置的阳光照射强度。强度0表示该位置点完全被树荫遮蔽，而强度1表示该位置点完全暴露在阳光下。2个游乐场日照强度的详细计算和分析将在数据分析和结果章节进一步解释。

4 激光雷达数据的采集与处理Lidar data collection and processing

5 光照强度分析Sunlight intensity analysis

2）（儿童）与自然的视觉距离是指从游乐场上的任何位置到最近的绿化地带的视觉距离，例如草坪、树木和树冠。较大的值表示目标儿童与自然元素之间的距离较长。如果目标儿童在有植被的地方或树冠下玩耍，那么其与自然的视觉距离计算为0。游乐场A的目标儿童与自然元素的视觉距离介乎0～5.18 m （0～17英尺）；游乐场B的目标儿童与自然元素的视觉距离在3.60～19.51 m（11.8～64.0英尺）之间。

3.4 数据收集日期和时间

实地数据收集于2018年3月14—17日进行。在这4天中的每一天里，一个专业的科研人员在3个时间段（11：30—12：00、14：00—14：30和18：00—18：30）对儿童的行为进行现场观察，总共产生了12个观察期。在观察期间，部分地区天气晴朗，室外最高温度约为12.20 ℃（54 ℉，2018年3月15日，星期四，中午），最低气温约为2.82 ℃（37 ℉，2018年3月16日，星期五，及3月17日，星期六，中午）。

3.5 抽样方法及目标儿童

目标儿童通过随机抽样来确定。第一个目标儿童是在观察员准备就绪后随机选择的。在每个儿童被观测6个观察区间后，观察员将刚刚进入主要游戏区域的儿童锁定为下一个目标儿童。观察员只是从游乐场的边界处进行被动观察，以避免干扰儿童游戏。最后，对A游乐场的94名儿童（564个观察区间）和B游乐场的91名儿童（546个观察区间）进行了记录，总共记录了1 110个游戏行为供进一步分析。

3.6 数据分析和结果

3.6.1 阳光和阴影分析

小飞机利用激光扫描仪将激光脉冲传送到地面，从地面读取反射或散射的激光，即为“光探测和测距技术”（Lidar）。通过该技术获得的数据可用来创建高质量的数字高程模型（Digital Elevation Models, DEMs），精准度可以在10 cm以内。激光雷达探测和测距技术以前曾应用于大规模的场地规划研究，如水文学、地形学和视域模拟[43-44]。为了更好地了解研究地点的阳光和阴影动态，读取了来自美国地质调查局国家地图观察器的激光雷达数据，以测量场地上遮阳结构的实际高度，如树木、屋顶结构和游戏器械等。图4展示了2个研究地点的激光雷达点云数据的处理结果，检测到了主要遮阳结构的高程数据，并通过截面图形进行可视化处理。在光探测和测距技术的帮助下，通过Rhino建立了场地的精确三维立体模型，包括每一个游乐场地的游戏设施、遮阳结构、植被、建筑物和地平面。这些三维模型作为地理坐标参考，并通过Grasshopper Ladybug②里的日照轨迹功能（sun path）合成仿真模型，同时在该模型中引入了标准的EnergyPlus天气数据（EPW格式），最终进行太阳辐射和遮阳研究。如图5所示，研究地点在每个观察期内，大约每0.1 m2（即每平方英尺）的光照强度数即可由此推算出来。

表1 用多分类逻辑回归分析探讨日照强度与自然元素的视觉距离和儿童的社交游戏行为的关系（N = 1 110）Tab. 1 Summary of multinomial logistic regression analysis of sunlight intensity and visual distance to nature to predict all children's social play behaviors (N = 1,110)

3.6.2 树木和社交游戏行为

案例研究Ⅰ共观察到1 110次儿童游戏行为，其中游乐场A观察到564次，游乐场B观察到546次。在所有观察中，5.0%（N=55）为非游戏行为，15.3%（N= 170）为单独游戏行为，41.6%（N= 462）为平行游戏行为，38.1%（N= 423）为群体游戏行为。统计分析使用了IBM SPSS统计软件。所有统计学检验均采用了传统的0.05统计学显著性标准。

多分类逻辑回归分析（multinomial logistic regression）用来评估儿童的社交游戏行为（结果变量）和2个预测变量（日照强度和与自然的视觉距离）之间的关系。在基本模型中添加这2个预测变量能显著提升模型的拟合性，χ²（6，N=1 110）= 188.182，p＜ 0.001。非游戏行为相对于群体游戏行为，当儿童与自然的视觉距离每增加一个单位时，在模型常数中保持其他变量不变的情况下，非游戏行为的多项对数概率会增加1.053个单位（p＜0.001）。相对于团体游戏，独自和平行游戏也有类似的效果：当增加儿童与自然的视觉距离时，表现出独自（OR= 1.04，p＜ 0.001）和平行游戏行为（OR= 1.05，p＜ 0.001）的可能性会显著增加。这些结果表明，当儿童在游乐场上与大自然有更多的视觉接触时，相比其他类型的游戏行为，他们更有可能参与集体游戏。日照强度只被认为是独自游戏行为的显著预测因子（OR= 0.441，p= 0.001）。与团体游戏行为相比，日照强度对于非游戏或平行游戏行为的预测效果不显著（表1）。3.6.3 树木和认知游戏行为

认知游戏行为的分布为：非游戏行为占4.5%（N= 50），功能性游戏为27.5%（N=305），建构性游戏为16.8%（N= 187），探索性游戏为15.6%（N= 173），戏剧化游戏为1.9%（N= 21），规则性游戏为30.5%（N=339），其他游戏行为占3.2%（N= 35）。多分类逻辑回归分析在此用来探讨儿童认知游戏行为与日照强度和与大自然的视觉距离这2个预测变量之间的关系。在基本模型中增加这2个预测变量显著改善了模型与数据之间的拟合度，χ²（12，N= 1 110）= 96.952，p＜0.001。

表2总结的结果显示，与规则性游戏相比，自然的视觉距离显著地预测了非游戏、功能性游戏、探索性游戏和戏剧性游戏行为。与规则性游戏相比，当孩子与自然的视觉距离增加时，他们表现出非游戏（OR= 1.03，p＜ 0.001）、功能性（OR= 1.019，p＜ 0.001）、探索性（OR=1.033，p＜ 0.001）和戏剧性游戏行为（OR= 1.039，p＜ 0.001）的可能性会显著增加。日照强度只能预测与游戏相关的功能性游戏行为：当孩子们在日照强度较强的地方玩耍时，他们更有可能参与功能性游戏，而不是参与规则性游戏（OR= 2.011，p= 0.001）。

3.6.4 树木和自由游戏行为

与自然的视觉距离可以显著预测儿童的自由游戏行为，χ²（2，N=1 110）=660.316，p＜0.001。当儿童与自然的视觉距离增加时，他们更有可能参与有引导的游戏/在游戏器材上玩耍，而不是自由地玩耍（OR= 1.617，p＜0.001）。

3.6.5 年龄和性别差异

根据Piaget的儿童认知发展阶段论，年龄和性别可能是潜在的调节因子，影响预测变量和结果变量之间的关系，幼儿和学龄前儿童的社交技能和团体游戏能力可能尚未完全建立[26]。因此，经过一系列的独立性卡方检测（Chi-square tests for independence）进一步探讨了儿童的年龄、性别和不同的游戏行为之间的相关性。结果表明，不同年龄组的儿童社交游戏行为差异显著，χ²（3，N=1 104）=61.807，p＜0.001。幼儿和学龄前儿童（5岁以下）表现出更多的非游戏和独自游戏行为，但较少参与群体游戏活动。6～12岁儿童显示出更多的群体游戏行为，却表现出较少的非游戏或独自游戏行为。年龄与儿童认知游戏行为显著相关χ²（6，N=1 104）= 85.264，p＜0.001。幼儿和学龄前儿童表现出明显更多的非游戏、探索性和戏剧性游戏行为，但较少参与规则性游戏。相比之下，6～12岁儿童能明显更多地参与到规则性游戏当中。同样，年龄与儿童的自由游戏行为显著相关，χ²（2，N=1 104）=22.422，p＜0.001。较年长的儿童显著表现出更多的自由游戏行为，而幼儿和学龄前儿童更有可能进行有引导的游戏。

表2 用多分类逻辑回归分析探讨阳光强度与自然元素的视觉距离和儿童的认知游戏行为的关系（N =1 110）Tab. 2 Summary of multinomial logistic regression analysis of sunlight intensity and visual distance to nature to predict children's cognitive play behaviors (N = 1,110)

性别与儿童社交游戏行为显著相关，χ²（3，N=1 110）=61.807，p＜0.001。男孩更多地参与团体游戏，而女孩更多地表现出非游戏、独自游戏和平行游戏行为。同样，性别与儿童认知游戏行为显著相关，χ²（6，N=1 110）=61.807，p＜0.001。男孩明显更多地参与规则性游戏。相比之下，女孩表现出更多的非游戏和探索性游戏行为。性别与儿童的功能性、建构性、戏剧性或占有性游戏行为之间没有显著的相关性。最后，性别与儿童的自由游戏行为显著相关，χ²（2，N=1 110）=12.123，p=0.002。男孩明显比女孩更喜欢自由玩耍，而女孩更喜欢在游乐场参与有引导的游戏。

4 案例研究Ⅱ

4.1 研究场所、数据和时间

研究Ⅱ是在美国俄克拉荷马州塔尔萨市的一个公共儿童主题公园“相聚之地”（The Gathering Place）进行的。“相聚之地”是一个占地约29.5 hm2（64英亩）的公共公园，坐落在阿肯色河沿岸，距塔尔萨市中心约3.22 km（2英里）。这个公园是由迈克尔·范瓦肯伯格景观设计事务所（Michael Van Valkenburgh Associates③）主持设计的。公园内经过精心设计的一系列游戏空间融入自然和地形之中[45]。“相聚之地”是一个享有盛誉的儿童公园，迄今为止赢得了许多奖项，包括《今日美国》的“最佳新旅游胜地”，以及《时代》杂志的“2019年全球最佳之地”等称号[46]。公园内有2个以梦幻仙境为主题的游乐场被选作案例研究Ⅱ的场地。这2个游乐场的设计兼顾了低挑战性的游戏项目，如滑梯、攀爬结构和游乐场，适合所有年龄组的儿童。这2个游乐场彼此相邻，并且都被丰富的自然环境所包围。在遮阴方面，游乐场A被四周的树冠遮蔽得很好，主游乐区内有一棵大型乔木提供遮阴（图6～7）。相比之下，游乐场B是一片开阔的空地，几乎没有树冠遮阴，主要的游乐设施完全暴露在阳光下（图8）。

研究Ⅱ的实地观察阶段于2019年4月27日进行。每个游乐场在5个时间段内观察儿童的游戏行为：10：15—10：45、11：00—11：30、13：30—14：00、14：15—14：45、15：00—15：30。在观测期间，天气晴朗，最高气温为25.0 ℃（77 ℉，在15：30达到），最低气温为15.6 ℃（60 ℉，在10：15达到）。

4.2 变量、数据收集工具和取样

案例研究Ⅱ采用与研究Ⅰ相同的工具和观察方案，以及相同取样策略，均为随机选择目标儿童。在数据收集上，研究Ⅱ与研究Ⅰ的2个细微差别是，GIS Cloud被传统的纸笔记录表所取代。而且，每个目标儿童最多被追踪15个观察区间，而不是6个。延长观察时间的目的是为了记录孩子们在每个游乐场上玩耍时间的差异。最终，在A游乐场观察了41名儿童（共268个观察区间），在B游乐场观察了53名儿童（共401个观察区间），总共观察到669个游戏行为来进行进一步分析。

6 案例研究Ⅱ游乐场A及遮阴情况Playground A in Study Ⅱ and the shading condition

7 案例研究Ⅱ位于游乐场A游戏区域的大型遮阴乔木A big shade tree at Playground A

8 案例研究Ⅱ游乐场B及日照情况Playground B in Study Ⅱ and the sunlight exposure

研究Ⅱ着重于3组结果变量，包括儿童的游戏行为（即社交游戏、认知游戏和自由游戏）、儿童在游戏过程中对树荫情况的选择（即完全荫蔽、部分遮阴和完全暴露在日照之下），以及儿童在游乐场的逗留时间。根据POS的要求，在任何一天最多只能收集同一个孩子5 min的行为数据，因此，任何在游乐场上停留超过5 min的目标儿童即不再被继续追踪。

4.3 数据分析和结果

4.3.1 游戏时间长度

皮尔森相关系数分析（Pearson's correlation analyses）用来探寻儿童的游戏时间的长短（单位为s）和其所在地的日照强度之间的关系。结果显示游戏时长与日照强度呈显著的负相关性，r= -0.459，N= 25，p=0.021，显著性特性标准为0.05。该结果表明儿童的游戏时间会随着游乐场光照强度的增加而缩短。这2个变量之间的关系呈中等强度[47]。研究结果表明有较多树荫的游乐场有延长儿童室外游戏时间的趋势。而对于幼儿/学龄前儿童，游戏时长与光照强度之间并无统计学意义上的相关性。

一系列的独立性卡方检测探讨了树荫与儿童不同游戏行为之间的关系，得到以下结果。首先，树荫与各年龄组儿童的社交游戏行为存在显著相关，χ²（6，N= 666）=125.685，p＜0.001。幼儿和学龄前儿童在完全荫蔽条件下玩耍时显示出更多的团体游戏 行 为，χ²（6，N= 461）=28.503，p＜0.001。然而6～12岁儿童则在部分遮阴条件下表现出明显的群体游戏行为，χ²（6，N= 205）=50.264，p＜0.001。其次，树荫与各年龄组儿童认知游戏行为显著相关，χ²（10，N=665）=168.642，p＜0.001。幼儿和学龄前儿童在完全荫蔽的环境下表现出更具建构性和戏剧性的游戏行为，在全阳光的环境下表现出更多的功能性游戏行为，χ²（10，N= 460）=32.933，p＜0.001。然而，在部分遮阴条件下χ²（8，N= 205）= 63.454，p＜0.001，6～12 岁儿童明显更多地参与到规则性游戏中。各年龄组儿童的自由游戏行为与树荫显著相关，χ²（2，N= 667）=44.671，p＜0.001。总的来说，各年龄段的儿童在完全荫蔽和部分遮阴条件下的自由游戏行为都要比在完全暴露在光照条件下表现得更为强烈。

性别与各种儿童游戏行为之间的相关性在统计学上显著。男孩的群体游戏行为显著多于女孩，但平行游戏行为显著少于女孩，χ²（3，N= 666）=17.324，p=0.001。男孩比女孩参与更多的规则性游戏，χ²（5，N= 665）= 15.945，p=0.007。男孩和女孩在自由玩耍行为上没有统计学上的显著差异。

5 讨论

世界卫生组织（WHO）将儿童健康定义为包括身体、精神、智力、社会、行为和情感健康的整体健康状态，而不仅仅是没有疾病。WHO的指导方针还强调了游戏和户外体育活动对儿童健康和福祉的重要性[48]。在全球化和快速城市化的时代，必须鼓励儿童更多地接触大自然，促进他们的游戏和户外活动，充分满足他们的发展需求。

笔者记录的2项研究是首批探讨城市化游乐场上的自然和景观树木如何影响儿童游戏行为的研究。对2个年龄组的儿童进行了观察，研究结果与一些经典理论的主张相一致，即儿童发展遵循等级阶段，从简单的感觉运动阶段到更高级的社会认知阶段[25]。幼儿和学龄前儿童表现出更多的独自游戏和功能性游戏行为，而6～12岁儿童能够更多地融入群体游戏中。研究结果表明，大自然和树荫的存在可以显著鼓励儿童的社交、认知和自由玩耍行为。总体而言，树荫良好的游乐场能促使各年龄段的儿童进行最高级别的游戏，无论是社交游戏还是规则性游戏。游乐场上有遮阴树的情况下，6～12岁儿童的玩耍时间显著增长。考虑到性别和年龄作为重要的调节因素，设计不同的游戏功能以适应不同的儿童群体，可以进一步提高游乐场的效率。

6 局限性和未来方向

首先，这项研究调查了美国亚热带湿润气候地区3个不同游乐场中孩子们的游戏行为。由于不同地区的地理、气候和社会人口状况各不相同，这些研究结果可能不适用于其他国家或地区的儿童游戏行为研究。这项研究重点调查了2个游戏季节，即春季和夏季。对于某些地理区域来说，严酷的冬季需要更多的阳光和热量，而这些游乐场上树木与儿童游戏行为之间的关系，以及他们的热舒适感知等，需要进一步的探讨。除了游乐场上树木和遮阴2个重要因素之外，还有其他类型的小气候因素（如地面辐射、湿度）也可能显著影响儿童在户外玩耍时的热舒适程度和行为结果[49]。

研究方法上，本研究采用观测量表等手段评估了游乐场内发生的游戏模式，这些方法最适合于评估人在不同环境中的瞬时行为。然而，研究表明，父母对风险和安全的态度，以及孩子自己对不同环境的看法，可能会影响他们选择在哪里和如何玩耍[50]。未来的研究应该使用综合性的研究方法，为树木在游乐场上的有益作用这一课题，以及树木影响儿童的游戏行为、健康和福祉的作用机制等研究，补充更多的循证设计依据。

正如英国国家儿童局（National Children's Bureau of the UK）发表的一份白皮书所总结的那样，自由玩耍有助于幼儿保持情绪平衡、身心健康、幸福感，以及在发展和学习的各个方面全面地掌握各项技能[51]。自由游戏对于塑造儿童早期发展至关重要；然而，在当今社会以人造游戏器材为主导的当代游乐场设计中，自由游戏或多或少被忽视了。当孩子们有足够的时间自由玩耍时，他们的游戏行为就会变得越来越复杂，对认知和社交的要求也越来越高。如何设计游乐场，将大自然中的各种零零碎碎融入其中，并提供足够的自由玩耍机会，需要在未来展开进一步的探索。

致谢：

1.直接物化成本部分。化肥和机械作业等“中间消耗费用”的增多，拉动了农业生产资料价格的大幅度上涨，现行农业保险甚至连最基本的农业生产直接物化成本都覆盖不了。

这项研究的部分经费由西弗吉尼亚大学2018—2019年Big 12 Fellowship资助。感谢西弗吉尼亚大学风景园林专业硕士生Udday Datta帮助制作文中的图表。感谢美国俄克拉荷马州立大学风景园林学助理教授张波博士协助这次实地考察，并感谢俄克拉荷马州立大学访问学生张冉在研究数据收集阶段提供的帮助。

注释：

① 在当今，美国1/5的儿童正被肥胖问题困扰。从美国疾病控制和预防中心（Centers for Disease Control and Prevention）获取了最新的数据和信息，网站：https://www.cdc.gov/obesity/childhood/。

② Grasshopper Ladybug网站：https://www.ladybug.tools/。③ Michael Van Valkenburgh Associates, Inc. 网站：https://www.mvvainc.com/。

图表来源：

图1～2由Udday Datta绘制；图3由玛丽亚米·马格莱克利兹提供；图4～5由宋阳提供；图6～8由姜珊摄影；表1～2由姜珊绘制。

Playground Trees' Impacts on Children's Play Behaviors and Well-being

Authors: (GEO) Mariami Maghlakelidze, JIANG Shan＊, SONG Yang, LI Dongying Translator: LI Xiaojie

1 Research Background

1.1 benefits of Contact with Nature on Child Development

Children in modern societies are experiencing a sharp decline of outdoor play and contact with nature, which leads to not only play deprivation but also nature deficit disorder, a concept describing the diminished use of the senses,attention difficulties, and higher rates of physical and emotional illnesses due to alienation from nature[1]. A growing number of scientific studies have revealed the negative health consequences of children's increasing indoor lifestyle, such as the rapid rise in childhood obesity rates and an elevated risk of several major non-communicable diseases①[2-3].The Biophilia Hypothesishighlights the health benefits of engaging with nature from evolutionary perspectives; it is the inherent human inclination to interact with nature and other living organisms[4]. Holding a biophilic perspective, Kahn and Kellert explored the psychological, intellectual,and sociocultural values in children being allowed to play in nature in their textChildren and Nature:Psychological, Sociocultural and Evolutionary Investigations[5]. Contact with nature is critical for nourishing children's holistic health status and facilitating their early childhood development.Indeed, research evidence supports a wide range of health benefits from contact with nature among the young generation, such as stress reduction,protection against eyesight problems, and the reduction of symptom severity in children with attention deficit hyperactivity disorder (ADHD)[6-9].Literature reviews of scientific studies that support the benefits of nature contact for children can be found in various articles and books[10-13].

1.2 Play and Play Behaviors

Play is an essential component of childhood development. Classic theories of play are composed of four branches that provide fundamental explanations about the reasons and functions of children's play behaviors, including 1) surplus energy[14], 2) recreational or relaxation[15],3) practice or pre-exercise[16-17], and 4) recapitulation theories[13,18]. Modern theories of play consider play to be a scheme that promotes cognition or symbolization, which offers understandings of the critical functions of play in children's development and rationalizes the power of the theoretical concepts as supported through empirical research,including psychoanalytic, arousal modulation,meta-communicative, and cognitive theories[13,19].Young children utilize play as the primary source to learn about the surrounding world, communicate their emotions and feelings, develop cognitive thinking and reasoning skills, and practice social and psychomotor skills[20]. Play is also involved in the development of personality. It encourages interpersonal relationships, stimulates creativity,adds to the joy of living, and advances learning[21-22].

There are many forms of play. When children play within a context that adults have set up (e.g., game rules or play equipment), they will likely experience a form of guided play[23];meanwhile, free play is an unstructured, voluntary,child-initiated activity that allows children to develop their imaginations while exploring and experiencing the world around them[24]. Children's play behaviors vary by factors such as age, gender,and children's developmental stages. According to Piaget's classification of three successive stages of child cognitive development, play behaviors can be classified according to the degree to which play remains purely sensorimotor or has some bearing on thought itself: 1) functional play,which involves the repetitive muscle movements with or without objects; 2) constructive play,which includes the manipulation of objects to construct or create something; 3) dramatic play, in which children substitute an imaginary situation to satisfy their personal wishes and needs; and 4) games with rules, which require the acceptance of prearranged rules and the adjustment to these rules[25-27]. In terms of the level of socialization,Parten discovered that social participation among preschoolers increased with the child's age in the following sequence: unoccupied behavior, solitary play, onlooker behavior, parallel play, associative play, and cooperative play[28]. Rubin developed the Play Observation Scale (POS) to describe the two play hierarchies (i.e., social and cognitive play behaviors, Tab. 1)[26]. POS has become the theoretical foundation and data collection tool of many established play studies[29-30].

1.3 Nature and Trees on Playgrounds

Based on play theories and the significant health benefits of engaging with nature, integrating natural elements into playground design has become a promising intervention to promote early childhood development. Such developmental benefits include: 1) significant contributions to children's cognitive and intellectual development in early childhood[31]; 2) enhancing children's physical activity levels and outdoor play time[32-34], and 3) encouraging young children's social interactions and emotional advancement during play[31,35-36].High levels of naturalness on playgrounds have been associated with restorative qualities[37]. Nature elements on outdoor playgrounds that children prefer include 1) water; 2) vegetation; 3) animals and creatures; 4) earth materials, sand and mud;5) natural color, texture, diversity, and change;6) seating opportunities under shade; 7) different levels of nooks and crannies (i.e., places that offer privacy and views); and 8) structures, equipment,and materials that can be changed in their imaginations, including plenty of loose parts[38].

Trees, the fundamental elements of any landscape design, can play a vital role at playgrounds for young children. For example, tree shade can raise playground safety by reducing surface temperatures and ultraviolet radiation exposure and enhancing the overall outdoor thermal comfort for children during play[39-40]. According to the research by American Academy of Dermatology,one in five Americans could develop some form of skin cancer during their lifetime, and five or more sunburns double the risk of developing skin cancer.Utilizing tree shade, designing play structures as a means for providing shade (e.g., elevated platforms with shaded space below), and creating more shade through manmade structures can help protect children's skin from sun exposure[41]. In a recent study, Anderson and colleagues audited 139 urban playgrounds and made 1,033 shade observations in Sydney, Australia; they found that the main activity areas at playgrounds in lower socioeconomic status areas had significantly less — about 34% less than the mean — tree shade coverage than higher socioeconomic status regions, which raised concerns about skin cancer prevention to evaluate trees on playgrounds within the socioeconomic context[42].

2 Research Questions and Methods

Substantial studies have highlighted the benefits of trees and shade in improving the thermal comfort, sun protection, and microclimate condition of playgrounds. However, there are comparatively insufficient explorations about the behavioral and social benefits of having trees and shade at playgrounds in cities. This article documented research findings from two recent studies about trees and shade at urban playgrounds and the impacts on children's play behaviors.Study I investigated how tree coverage and shade affect young children's play behaviors at two manufactured playgrounds in the city center of Washington, DC. Study II investigated how the tree coverage, shade, and other natural materials impact young children's outdoor play time, play behaviors,and activity patterns at two playgrounds within a themed children's park in the urbanized area of Tulsa, Oklahoma, in the United States.

Both studies utilized direct observation as the primary method. Young children's play behaviors were observed and mapped according to Rubin's POS[26]. Study I involved behavior mapping through the Geographic Information System (GIS) Cloud program. Additional variables in Study I included the two playgrounds' hourly sunlight intensity derived from the Lidar data and EnergyPlus weather data, as well as one's visual access to nature. Study II involved the real-time tracking of children's play behaviors onsite through the traditional paper-and-pencil technique; one's length of stay at the playground and the sun exposure condition were also recorded.

3 Study I

3.1 Study Sites and Characteristics

Two outdoor playgrounds in the city center of Washington, DC, were selected for Study I(Fig. 1-2). Playground A and B are located in the northeastern part of the city, approximately 4.8 km (3 miles) apart. Both playgrounds were established in 2013, and both the spatial design and play equipment design were created by the same playground design firm based in Minnesota in the United States. The design concept of both playgrounds represents a collaboration among mathematics, art, architecture, science, and nature. Both playgrounds offer play equipment to accommodate two age groups: toddlers/preschoolers (2-5 years old) and young children (6-12 years old). Play equipment on both sites include a seesaw, a sensory play center, shelters/huts, a bridge playset, a jungle gym playset, and swings.A nature-inspired color palette was adopted in the playscape design, including denim, lagoon, leaf,lemon, pine green, and tan colors. Playground A includes canopy trees that provide natural shade for the main play zones of the playground. Children have direct visual and physical access to several small pieces of lawns and lower shrubs from the main play areas. In contrast, although there are a few trees planted along the perimeter of Playground B, the main play areas are not shaded by any tree canopies.

3.2 Instrument and Observation Protocol

Data collection was guided by the POS[26],which requires the observer to observe the target child for a 10-second interval, then spend the next 5-10 seconds coding the predominant behavior and recording the behavior on a coding sheet. Thus,it will take 1.5-2 minutes to obtain 1 minute of recorded observations. In order to obtain a valid measure of the child's general play styles, the POS recommends that only up to 5 minutes of the same child's behavior be recorded on any given day. The scale also recommends gathering a minimum of 15 minutes of POS data for analysis and result interpretation. Therefore, Study I observed each target child for six intervals, and each observational period lasted for 30 minutes. Following the POS,during each 10-second interval, only one behavior is coded. If more than one behavior occurs during a 10-second interval, the behavior expressed for the majority of the time sample is coded. If behaviors are of the same length, the observer should code the child's behavior following a hierarchy: any group behavior supersedes all other behaviors(i.e., group play > parallel play > solitary play). For cognitive play behaviors, games with rules override other constructive or functional play behaviors (i.e.,games with rules > dramatic play > exploration >constructive play > functional play)[26].

In addition to the social and cognitive play behaviors described in the POS, whether the child played on equipment or not was also recorded. In Study I, playing on any equipment was defined as“guided play” while those who engaged in selforiented play and games away from the equipment zones were defined as engaging in “free play”.

The GIS Cloud (GIS Cloud Inc., Zagreb,Croatia) is a web-based, real-time mapping tool based on the geographic information system. The observer was able to map the location of the target child within any 10-seceond interval and record the child's predominant behavior through a predeveloped coding sheet embedded in GIS Cloud.Fig. 3 depicts the user interface of the mapping and coding function of GIS Cloud.

3.3 Additional Variables

Two additional variables indicating the sun/shade condition and one's access to nature were also measured/retrieved, including the following:

1) Sunlight intensity, which refers to the intensity of sunlight exposure at any location mapped on the playground within an observational period. A 0 intensity indicated that the location was fully shaded, while a 1.0 intensity indicated that the location was fully exposed to sunlight. Detailed calculation and analysis of sunlight intensities for the two playgrounds will be further explained in the Data Analysis and Results section.

2) Visual distance to nature refers to the visual distance from any location mapped on the playground to the nearest greenery, such as the lawn,plants, and tree canopy. A larger value indicated the longer distance between a target child and nature.If the target child was playing in a vegetated area or right under a tree canopy, then the visual distance to nature was calculated as 0. For Playground A, the target children's visual distances to nature ranged from 0-5.18 m (0-17'); for Playground B, the target children's visual distances ranged from 3.60-19.51 m(11.8'-64.0').

3.4 Data Collection Date and Time

Field data were collected on March 14-17, 2018. On each of the four days, a trained researcher conducted onsite observations of children's behaviors during three time periods (i.e.,11:30-12:00, 14:00-14:30, and 18:00-18:30), which resulted in 12 observational periods in total. During the observational periods, the weather onsite was sunny and clear, with the highest temperature being about 12.20° C (54° F, Thursday, March 15, 2018,at noon) and the lowest temperature being about 2.82° C (37° F, Friday, March 16 and Saturday,March 17, 2018, at 18:00).

3.5 Sampling and Target Children

Target children were randomly sampled.The first child was randomly selected right after the observer set up. After each child was observed for six intervals, the observer selected the next child based on whoever had just entered into the main play zone. The observer only conducted passive observations from the boundary of the playground to avoid any distraction to children's play. Ultimately, 94 children (564 observations) at Playground A and 91 children (546 observations)at Playground B were observed, which resulted in a total of 1,110 play behaviors for further analysis.

3.6 Data Analysis and Results

3.6.1 Sun and Shade Analysis

Collected by a small plane using a laser scanner that transmits laser pulses to the ground surface and reads reflected or scattered laser back from the ground, Lidar, which stands for light detection and ranging, can create high quality digital elevation models (DEMs) with an accuracy as good as 10 cm. Lidar has been previously applied in large-scale planning studies, such as hydrological,topographical, and viewshed modeling[43-44]. To better understand the sun and shade dynamics of the study sites, Lidar data were acquired from the National Map Viewer of USGS to measure the actual height of the shading structures, such as trees, roof structures, and play equipment. Fig.4 depicts the data processing results of the Lidar point cloud for the two study sites; all elevation data of the key shading structures were detected and can be visualized through cross-section graphics.With the assistance of the Lidar measurements, an accurate 3D Rhino model including play structures,shading structures, vegetation, buildings, and ground levels was built for each of the playground sites. These 3D models were georeferenced and integrated with a Sun Path simulation model named Grasshopper Ladybug②, which imports standard EnergyPlus Weather files (.EPW) into Rhino and conducts solar radiation and shading studies. As shown in Fig. 5, the total direct sunlight hours of each square foot of the study site was calculated for the observational periods.

3.6.2 Trees and Social Play Behaviors

A total of 1,110 observations were obtained,including 564 observations from Playground A and 546 observations from Playground B. Out of all observations, 5.0% (N= 55) were non-play behaviors, 15.3% (N= 170) displayed solitary play,41.6% (N= 462) displayed parallel play, and 38.1%(N= 423) were group play behaviors. The statistical analysis was conducted using IBM SPSS Statistics(IBM Corp., Armonk, NY, USA) program. The traditional .05 criterion of statistical significance was employed for all tests.

A multinomial logistic regression was conducted to assess the relationship between children's social play behaviors (outcome variable)and two predictor variables: sunlight intensity and visual distance to nature. Adding the two predictors to a base model that contained only the intercept significantly improved the fit,χ² (6,N= 1,110) =188.182,p＜0.001. For non-play relative to group play behaviors, when a child's visual distance to nature was increased by one unit, the multinomial log-odds of displaying non-play behaviors would be expected to increase by 1.053 unit while holding other variables in the model constant (p＜0.001).Similar effects were identified for solitary and parallel play relative to group play: When increasing a child's visual distance to nature, the likelihood of displaying solitary (OR= 1.04,p＜0.001) and parallel play behaviors (OR= 1.05,p＜0.001) would increase significantly. These results revealed that,when children have more visual contact with nature at a playground, they are more likely to engage in group play than other types of play behaviors.Sunlight intensity was found as a significant predictor for solitary play behaviors only (OR=0.441,p= 0.001) The predicting effects were not significant for the non-play or parallel play relative to group play behaviors (Tab. 1).

3.6.3 Trees and Cognitive Play Behaviors

Out of all 1,100 observations, the distribution of cognitive play behaviors were: 4.5% (N= 50) for nonplay behaviors, 27.5% (N= 305) for functional play,16.8% (N= 187) for constructive play, 15.6% (N=173) for explorative play, 1.9% (N= 21) for dramatic play, 30.5% (N= 339) for games with rules, and 3.2% (N= 35) for other occupied play behaviors.A multinomial logistic regression was conducted to explore the relationship between children's cognitive play behaviors and two predictor variables: sunlight intensity and visual distance to nature. The addition of both predictors to a model that contained only the intercept significantly improved the fit between model and data,χ² (12,N= 1,110) = 96.952,p＜0.001.

The results summarized in Tab. 2 revealed that visual distance to nature significantly predicted non-play, functional play, explorative play, and dramatic play behaviors relative to games with rules. Compared with games with rules, when children's visual distance to nature increased, their likelihood of displaying non-play(OR= 1.03,p＜0.001), functional (OR= 1.019,p＜0.001),explorative (OR= 1.033,p＜0.001), and dramatic play behaviors (OR= 1.039,p＜0.001) were expected to increase significantly. Sunlight intensity predicted only functional play behaviors relative to games: When children played in an area with stronger sunlight intensity, they were more likely to engage in functional play than games with rules(OR= 2.011,p=0.001).

3.6.4 Trees and Free Play Behaviors

Visual distance to nature significantly predicted children's free play behaviors,χ² (2,N=1,110) = 660.316,p＜0001. When children's visual distance to nature increased, they were more likely to involve in guided play/play on equipment than playing freely (OR= 1.617,p＜0.001).

3.6.5 Age and Gender Differences

According to Piaget's children developmental stage, age, and gender could be potential moderators that affected the direction of the relationship between predictors and outcome variables;toddlers' and preschoolers' social skills and group play capabilities may have not been fully established yet[25]. Therefore, additional associations among children's age group, gender, and different play behaviors were explored through multiple Chisquare tests for independence. The results revealed that age group was significantly associated with different social play behaviors,χ² (3,N= 1,104)= 61.807,p＜0.001. Toddlers and preschoolers(younger than 5 years old) displayed significantly more non-play and solitary play behaviors but were less involved in group play activities. Young children(6-12 years old) displayed significantly more group play behaviors but fewer non-play or solitary play behaviors. Age group was also significantly associated with children's cognitive play behaviors,χ² (6,N= 1,104) = 85.264,p＜0.001. Toddlers and preschoolers displayed significantly more non-play,explorative, and dramatic play behaviors, but were less involved in games with rules. In contrast, young children were significantly more involved in games with rules. Likewise, age group was significantly associated with children's free play behaviors,χ²(2,N= 1,104) = 22.422,p＜0.001. Young children significantly demonstrated more free play patterns,while toddlers and preschoolers were more likely to involve in guided play.

Gender was found to be significantly associated with children's social play behaviors,χ² (3,N=1,110) = 61.807,p＜0.001. Boys were significantly more involved in group play while girls displayed more non-play, solitary play, and parallel play behaviors. Likewise, gender significantly associated with children's cognitive play behaviors,χ² (6,N=1,110) = 61.807,p＜0.001. Boys were significantly more involved in games with rules. In contrast,girls displayed significantly more non-play and explorative play behaviors than boys. There were no significant associations between gender and children's functional, constructive, dramatic, or occupied play behaviors. Finally, gender was found significantly associated with children's free play behaviors,χ² (2,N= 1,110) = 12.123,p= 0.002.Boys were significantly more involved in free play than girls, who were more likely to involved in guided play at the playground.

4 Study II

4.1 Study Sites, Data, and Time

Study II was conducted at the Gathering Place, a public children's park in Tulsa, Oklahoma,in the United States. The Gathering Place is a about 29.5 hm2(64-acre) park situated along the Arkansas River and about 3.22 km (2 miles) from downtown Tulsa. The park was designed by Michael Van Valkenburgh Associates③, where a series of richly programmed play spaces were blended into nature and topography[45]. The Gathering Place is an iconic children's park that has won numerous awards,includingUSA Today's “Best New Attraction” in the United States andTIMEmagazine's World's Greatest Places 2019[46]. Within the park, two fairyland-themed playgrounds were selected for Study II. Both playgrounds were designed to accommodate low-challenging play opportunities,such as slides, climbing structures, and playhouses,which were appropriate for all age groups of children. The two playgrounds were adjacent to each other, and both were surrounded by a rich natural context. In terms of shading conditions,Playground A was well-shaded by surrounding tree canopies, and there was a large shade tree in the main play zone (Fig. 6-7). In contrast, Playground B was an open field with few tree canopies for shade,and the major play features were fully exposed to sunlight (Fig. 8).

The onsite observation phase of Study II was conducted on April 27, 2019. Five observational periods were conducted on each playground: 10:15-10:45, 11:00-11:30, 13:30-14:00, 14:15-14:45, and 15:00-15:30. During the observational periods the weather was sunny and clear onsite, with the highest temperature being 25 °C (77° F) at 15:30 and the lowest temperature being about 15.6 °C(60° F) at 10:15.

4.2 Variables, Instrument, and Sampling

Study II utilized the same instrument and observation protocol as Study I. Target children were selected at random using the same strategy as Study I. Two minor differences in the data collection in Study II were that GIS Cloud was replaced by the traditional paper-and-pencil coding sheets and each target child was tracked for a maximum of 15 intervals (instead of six intervals).The purpose of the extended observational interval was to capture any differences in play time the children spent on each playground. Ultimately, 41 children (268 observations) at Playground A and 53 children (401 observations) at Playground B were observed, which resulted in a total of 669 play behaviors for further analysis.

Study II focused on three sets of outcome variables, including children's play behaviors (i.e.,social play, cognitive play, and free play), children's choice of sun/shade condition during play (i.e., full shade, partial shade, and full sun condition), and children's entire length of stay at the playground.To follow the POS's requirement of collecting only up to 5 minutes of the same child's behavior on any given day, any children staying at the playground for longer than 5 minutes were no longer tracked.

4.3 Data Analysis and Results

4.3.1 Length of Playtime

Pearson's correlation analyses were conducted to explore the length of children's playtime (unit in seconds) in different sun/shade conditions at the two playgrounds. There was a significant, negative correlation between the children's playtime and the sunlight condition,r= -0.459,N= 25,p= 0.021(at 0.05 significance level), which indicated that young children (6-12 years old) spent significantly less time playing in a play area with higher sunlight intensity at the playground; the strength of the relationship was medium[47]. This indicated that a playground with more tree shades tended to prolong young children's outdoor playtime.Differences in the length of playtime for toddlers/preschoolers among different sun/shade conditions were not statistically significant.

4.3.2 Tree Shade and Different Play Behaviors

Several Chi-square tests for independence were conducted to explore the correlations between tree shade and children's different play behaviors.First, tree shade was found to be significantly correlated to the all ages group of children's social play behaviors,χ² (6,N= 666) = 125.685,p＜0.001. Toddlers and preschoolers displayed significantly more group play behaviors when playing in the full shade condition,χ² (6,N=461) = 28.503,p＜0.001. However, young children displayed significantly more group play behaviors in a partial shade condition,χ² (6,N= 205) =50.264,p＜0.001. Second, tree shade was found to be significantly correlated to the all ages group of children's cognitive play behaviors,χ² (10,N= 665) = 168.642,p＜0.001. Toddlers and preschoolers displayed significantly more constructive and dramatic play behaviors when playing in a full shade condition, and they tended to display more functional play behaviors in a full sun condition,χ² (10,N= 460) = 32.933,p＜0.001.However, young children were significantly more involved in games with rules in a partial shade conditionχ² (8,N= 205) = 63.454,p＜0.001.Finally, tree shade was found to be significantly correlated to the all ages group of children's free play behaviors,χ² (2,N= 667) = 44.671,p＜0.001.Generally, children displayed stronger free play behaviors in both full shade and partial shade conditions than the full sun condition.

The correlations between gender and children's various play behaviors were found to be statistically significant. Boys displayed significantly more group play but fewer parallel play behaviors than girls,χ² (3,N= 666) = 17.324,p= 0.001.Boys displayed significantly more games with rules than girls,χ² (5,N= 665) = 15.945,p= 0.007.There was no statistically significant difference between boys and girls on free play behaviors.

5 Discussion

The World Health Organization (WHO)defines child health as a holistic state of well-being that includes the physical, mental, intellectual,social, behavioral, and emotional well-being, not merely the absence of disease or infirmity. The WHO's guidelines also emphasize the importance of play and outdoor physical activities on children's health and well-being[48]. In the era of globalization and rapid urbanization, it is imperative to support children's contact with nature and promote their play and outdoor activities to achieve their full developmental potential.

The two studies documented in this article were among the first to explore how nature and landscape trees on urbanized playgrounds could impact children's play behaviors. Two age groups of children were studied, and the findings were consistent with the assertions of some classic theories that early childhood development follows hierarchical stages — from the simple,sensorimotor stage to a more advanced, socialcognitive stage[25]. Toddlers and preschoolers displayed more solitary and functional play while young children were able to play more games in groups. The findings revealed that the presence of nature and tree shade could significantly encourage children's social, cognitive, and free play behaviors.Generally, the playground with good tree shade encouraged the highest level of play behaviors across all age groups of children — both social play and games with rules. The presence of shade trees at a playground was significantly associated with longer playtime for children, particularly those 6 to 12 years old. Considering gender and age as significant moderators, designing play features to accommodate different groups of children could further improve the efficiency of the playground.

6 Limitations and Future Directions

First, this study examined children's play behaviors at three different playgrounds in the humid subtropical climate zone in the United States.As the geographic, climatic, and socio-demographic conditions vary across different regions, the findings may not be generalizable to children's play in other countries or regions. This study examined two play seasons (i.e., spring and summer), and for other geolocations where the excessive sunlight and heat are desirable in the harsh winter, the relationship between trees, children's play behaviors,and their perceived thermal comfort needs further exploration. Aside from the importance of trees and shade at playgrounds, there are other types of microclimate factors (e.g., terrestrial radiation,humidity) that could significantly impact children's thermal comfort levels and their behavioral outcomes during outdoor play[49].

Methodologically, this study employed auditing tools to assess the play patterns that occur at playgrounds, which are best suited to assess instantaneous behaviors with respect to different environmental exposures. However, studies have shown that parents' attitudes towards risks and safety, as well as children's own perceptions of different environments, may influence where and how they choose to play[50]. Future studies should use diverse methods to supplement the evidence of playground trees' benefits and mechanisms that may impact children's play behaviors and well-being.

As summarized in a white paper published by the National Children's Bureau of the UK, free play helps young children maintain emotional balance,physical and mental health, and well-being as well as master skills across all aspects of development and learning[51]. Free play is crucial in shaping early childhood development; however, it has been somewhat neglected in the design of contemporary playgrounds dominated by manufactured play equipment. When children are given enough time to play freely, their play grows in complexity and becomes more cognitively and socially demanding.How to design playgrounds to incorporate the broad variety of loose parts from nature and offer sufficient opportunities for free play require further explorations in the future.

Acknowledgments:

The study was partially funded by West Virginia University’s 2018-19 Big 12 Fellowship. We thank Udday Datta, MLA Candidate, West Virginia University for developing the figures in the article. We thank Dr. Zhang Bo, assistant professor of Landscape Architecture at Oklahoma State University, for hosting the site visit, and Zhang Ran, MLA,visiting student at Oklahoma State University, for the assistance during the data collection phase of the study.

Notes:

① Obesity now affects 1 in 5 children and adolescents in the United States. Up-to-date information can be found at the Centers for Disease Control and Prevention (https://www.cdc.gov/obesity/childhood/).

② Grasshopper Ladybug (https://www.ladybug.tools/).

③ Michael Van Valkenburgh Associates, Inc. (https://www.mvvainc.com/).

Sources of Figures and Tables:

Fig. 1-2 © Udday Datta; Fig. 3 © Mariami Maghlakelidze;Fig. 4-5 © SONG Yang; Fig. 6-8 © JIANG Shan; Tab. 1-2 ©JIANG Shan.

(Editor / WANG Yilan)