中国海绵城市—一座巨型绿色屋顶实验室

著：（加）丹尼尔·勒尔 译：鲍沁星 校：王思思

自2015年海绵城市在中国16个城市开展以来，低影响开发（LID）已取得了广泛的成就[1]（图1）。LID的目标是模拟场地在未受人类干涉的自然状况下雨水径流的运作模式。LID措施包括：绿色屋顶、生物滞留设施、透水铺装、生物洼地、储存/调节塘、雨水收集和储存装置（图2）。戴莉萍在最近的一篇文章中详细探讨了这一中国新型城市雨洪治理项目在管理结构方面的利弊及面临的融资挑战[2]。几个世纪以来，中国各级政府都试图通过大坝等结构性措施和政策等非结构措施来降低城市洪涝灾害[3]。邵薇薇在其海绵城市研究的文章中列举了过去8年间的重大洪灾以及因此造成的人员伤亡和经济损失，由此强调构建海绵城市数据计划对于改善城市设计和建设的重要性[4]。文中还指出“海绵城市的建设是一项非常复杂的系统工程任务”[4]，而海绵城市数据的收集对于提高防汛效率有重要作用，有利于海绵城市的成功建设。

对中国来说，治理洪水并非话题。事实上，中国人民5 000多年来不得不创造性地想出各种防洪措施来应对频发的洪灾[3,5]。吴庆洲在1989年的文章中不仅详细解释了中国的防洪历史、治理结构和政策，还介绍了古代中国城市规划政策中关于开展防洪实验的天才技术智慧和决心[5]。中国悠久的防洪历史与许多城市的地理位置有一定关系，它们往往建设于沿海地区或是广阔的河流三角洲地带，这些地区由于季风性气候定期受到洪水的侵扰。这些区域有着大量居民，常因季风而饱受洪灾之苦（图3）。而我们知道，中国的降水量，从东南向西北地区是逐步减少的[3]。

当下城市面临的众多挑战使得雨洪问题进一步复杂化，这直接促使了2015年海绵城市快速地启动。随着中国人口增长，越来越多的农业用地用于城市化，改变了原有的土地利用模式[3]。经济繁荣带动了如火如荼的新城建设，成千上万的人们涌入城市的“旷地”。城市中大量不透水下垫面，阻碍了土壤对雨水的渗透吸收。另外气候变化使得降雨情况更加难以预测[3]。短时间内的高强降水将会提高雨洪的峰值流速，导致雨洪径流不可控，加重土壤、河道、海洋的水污染[3]。

1 海绵城市示例Sponge City example

2 绿色屋顶： LID措施及LID系统组成部分Living roofs: a tool and systems component of LID

但是，对于拥有如此漫长防洪治洪历史的中国来说，怎样才能积极有效地解决这些问题呢？中国人适水而生。生活在乌镇、南浔、同里等江南水乡的人们几个世纪以来都与当地的河湖和谐相处（图4）[6]。如果公众和决策者能够学习水乡人看待水的态度，不将城市中的雨水径流看成是一种威胁，而视为一种需要保护的珍贵资源，并通过设计使之成为LID这一雨水管理体系中的一部分，不失为解决上述雨洪问题的一个可能性方案。这些LID系统能够减少径流量、削减洪峰、实现源头控制和雨水就地收集利用。“简单式”绿色屋顶就是这样一种有效的LID措施，能在城市雨水管理方面发挥重要作用，尤其是市中心区域，这些地方由于缺少透水面层，往往无法管理地表的雨水径流。

景观设计师在利用LID进行雨洪管理这项任务上处于有利地位。他们不仅在实施阶段，还能在LID倡导和可视化阶段发挥主导优势。LID系统的成功创建需要景观设计师和工程师从项目伊始就进行合作，前者负责LID系统的设计并对规模进行预估，后者则负责计算确定最终规模。景观设计师不仅能够从场地、城市和区域3个层面上对场地及其周围复杂互联的环境进行全面解读，而且掌握需要实施LID的其他因素的完整知识，包括当地气候、土壤排水能力、场地总体情况、分级和排水情况、场地现状。他们还懂得适时适地选择相应的LID措施以及实现最优效益的LID组合方式。可以说，如果能在城市整体规划阶段伊始就进行LID雨水管理体系规划，而不是作为一种事后补救措施，那么就能在一开始确定更为合理的建筑规划，有利于实现雨洪的就地管理。随着中国先进的海绵城市规划的就位，应用“简单式”绿色屋顶管理城市雨洪政策的相应实施，中国城市有望很快引领世界LID设计研究的前沿。

1 国际性“简单式”绿色屋顶

在过去几年中，有关“简单式”绿色屋顶的雨洪调控能力方面的研究可谓是相当科学和成功。“简单式”绿色屋顶是指：“单独或组合式的单层或多层人工绿色屋顶系统（通常包括：排水层，栽培介质层和植物层）。‘简单式’绿色屋顶的栽培介质在25～150mm之间，表面可以是水平或具有一定低缓地形的，植物层通常栽植矮生植物”[7]。本文作者提及的绿色屋顶指“简单式”绿色屋顶，因为“花园式”屋顶绿化建造和维护都十分复杂，不符合城市雨洪管理经济、易行的要求。

3中国2018年洪水灾害案例Flood example China 2018

4同里水乡Water City Tongli

绿色屋顶削减屋面径流的能力主要与当地气候、蒸散率（ET）、工程土壤介质、屋顶组件和植物有关。因此屋顶植物的选择和种植环境的构建应该适应当地气候条件，保证绿色屋顶成功运作。伊丽莎白·法斯曼·贝克说：“就绿色屋顶的雨洪管理功能而言，栽培介质层可能是最重要的组件—因为栽培介质本身就像海绵一样。[7]”单靠绿色屋顶并不能完全成功解决雨洪管理问题。在极端气候下，比如在夏季漫长的干旱期，屋顶的植物需要额外的浇灌，而普通绿色屋顶无法提供贮水和水资源再分配所需的能量。又比如，在屋顶没有LID雨水管理系统的情况下，就必须在建筑内部设置额外的蓄水池来滞留极端降水事件时屋面的雨水径流。也就是说，普通的绿色屋顶只能应对常规的降水事件。谢菲尔德大学副教授弗吉尼亚·斯图文，在她2015年的论文中称，在她的实验场地“在常规暴雨情况下，降雨截流效益良好……但是没有一处能够达到暴雨强度为2L/s/hm2时峰值截流能力的要求（等同于绿地雨水径流截留能力），来应对更极端的降水事件（例如，实测的日降水量P＞10mm的情况或者模拟30年一遇的情况）”[8]。由此看出，绿色屋顶只是一种阶段性措施，为了有效管理甚至净化雨水，需要结合其他LID措施协同应用，如生物滞留设施、洼地、储存/调节塘等，形成完整的雨水治理链。当“简单式”绿色屋顶的雨水径流量达到田间持水量时，便失去了雨水调控能力，变得与普通屋顶无异，这时候协同的LID措施能够帮助有效处理屋面的峰值雨水径流。出于对绿色屋顶雨水管理能力和成本的考量，中国的政府和设计实践界尚未完全接受绿色屋顶理念。但是有很多学者坚信绿色屋顶的雨洪治理效益，比如同济大学董楠楠副教授和骆天庆副教授，他们是中国绿色屋顶研究的开拓者，现有的相关研究大部分都来自他们。

在过去几年中，一些重要的国际性研究已经证明了绿色屋顶的雨洪管理效益。例如，斯图文和她的团队多年来一直在谢菲尔德大学实验室测试不同基质媒介（介质）和植被组合的水文性能，并在2015年的一篇最新论文中得出结论，“研究表明，市面上最典型的‘简单式’绿色屋顶的结构，即砖底基质层和景天类植物层的组合方式在雨水截流和滞留方面具有最全面优异的水文性能。然而，应该指出的是，浅层的‘简单式’绿色屋顶系统需要与位于LID治理链下游的截流和滞留措施相结合，形成更为整体的可持续城市排水系统（SUDS），这样即便最大的暴雨事件下的洪水风险也能降低”[8]。她同时指出：“长期植被覆盖、分级良好、基质渗透性较弱的绿色屋顶的雨水截流能力优于配备大孔径渗透性强基质的屋顶。[8]”

在德国，景天类植物作为屋顶绿化材料中的“主力军”已经成功使用40余年（图5）[9]。斯图文的新研究再次证实，景天类植物与渗透性较弱基质的组合方式在雨水截流方面卓有成效。澳大利亚的另一支团队在评述了不同植物组合方式和用水模式对雨水截流能力的影响后得出结论：“与我们的假设相反，植物层种植用水可塑性植物的绿色屋顶的雨水截流效果并不比栽植如乙女心（Sedum pachyphyllum）之类的保守型用水的多肉植物的屋顶好。另外，多种植物混植的屋顶绿化也没能比单一植物截流更多的雨水。这是因为多种植物混植的屋顶绿化虽然提高了屋面植物层的生物总量和生产力，但屋面的整体雨水调蓄性能会受到其关键物种—弯药百合（Stypandra glauca）的不利影响。其根系很容易穿过基质形成优先径流通道，造成雨水流失，从而降低基质的储水量、蒸散量和雨水截流量。因此为了优化屋面的雨水截流能力，提高植物存活率，屋顶绿化的植物选择不能仅考虑植物用水模式或物种多样性，还应考虑植物的根系性状，因为植物根系形成优先径流通道造成的影响远远超过了用水可塑性植物的用水特点带来的好处。[10]”

上述两个例子展现了现有绿色屋顶雨水截流效益和局限性研究中的研究深度、细节和严谨性。而中国海绵城市建设为这些研究提供了契机。如上述实验，现有的大多数实验都是在屋顶或同等规模下可控的小型实验地完成的。中国现在有机会借海绵城市建设之东风将实验扩展到城市规模，包括对绿色屋顶从建成之初就开始进行全面监测。这不仅能帮助中国人延续自古以来前沿的防洪研究传统，而且有助于保护城市免遭重大洪水的侵害，引导公众将雨水视为城市设计中的积极因素。绿色屋顶在LID雨水处理链中属于前期措施，如上文所述，单独使用绿色屋顶并不能可靠有效地应对大型暴雨事件。因此，在城市尺度下设计和构建基于LID的绿色屋顶至关重要。

绿色屋顶对LID雨洪管理系统的好处：1）通过植物和土壤蒸发减少进入市政雨水管道的雨水径流。2）减少因应对气候变化所需的市政雨水管道的升级。3）在城市密集区实现雨水源头净化和地下水补充。4）形成构筑物与周围环境间连续的生态栖息地。5）丰富“视觉体验”，有利于人类健康。6）改善相邻建筑物周围的小气候。

2 设计师的责任

景观设计师在进行绿色屋顶和LID设计时需要承担一定的责任。一旦出了问题，设计单位乃至整个行业的经济和名誉都会遭受重大损失，后果十分严重。因此，在设计和施工监理的过程中需要始终保持严谨和专业。景观设计师们应该在设计过程中早早行动，在所有项目的城市规划阶段就提倡绿色屋顶建设。事实上，他们应是同城市规划师和建筑师一起在项目现场的第一批设计师。海绵城市理念为景观设计师成为项目主导提供了一个绝佳机会。LID是项目启动时考虑的第一个问题。戴莉萍在其文章中介绍了从中央政府到地方当局自上而下式的海绵城市管理机制[2]。她强调做好城市防洪工作不仅仅是技术问题，而且还是一个管理问题—尤其是在海绵城市规划之初[2]。设计师们应该试着充分认识理解项目所在地的管理结构，在此基础上提出与之适应的LID策略。戴莉萍将中国海绵城市的管理结构分成四类模式：自发型；规制型；供给型；赋能型[2]。这样的管理结构为设计师提供了一个框架，指导他们如何参与到上述4类管理模式中。

自发型管理是海绵城市的自我管理，比如像戴莉萍在论文中解读武汉案例那样分析建成项目[2]。景观设计师可以通过LID可视化方式强调绿色屋顶的雨洪管理效益，通过在线动画向公众生动形象地解释LID整体系统，在TED上公开演讲来宣扬LID的重要性，还可以协同当地居民和政府共同公开演示LID这一新措施。目前中国海绵城市建设的政策中并没有强制要求公众参与[2]，这与中国的传统文化的价值观念有关—国家利益高于集体，个人利益在最后。这些观念深植在中国人的规划理念里，当下依然盛行，在今后的LID设计中仍需要引起重视[2]。景观设计师需要不断学习LID技术和专业知识，还要通过出版著作、研讨、会议、公开课、展览以及社交媒体等形式向公众宣传相关知识，想方设法提高公众参与，与公众分享他们的专业技能和想法。规制型管理是运用传统的法律法规进行管理的模式[2]。景观设计师们可以提倡将LID措施作为一种强制性的法规纳入城市治水法律体系之中，强调绿色屋顶在城市尤其是高密度城市区域雨洪管理的效益。通过回顾现有最新的国内外业内研究，可以肯定的是，技术性条例需要根据城市当地的地理情况特点制定[2]。在供给型模式下，设计师可以向当局主动提供LID的资源、工具包（比如LID计算器等）、国际性的技术知识。甲方、政府人员、施工人员和研究人员是LID的共同决策者，景观设计师在项目规划之初“教育”他们深入理解项目位置、雨洪现状及项目背景是十分重要的。可以说，景观设计师充当了研究学者和甲方、当局和施工安装人员之间的桥梁。由于私人投资者对政府不信任，在中国实现赋能型管理是十分困难的[2]。景观设计师可以提供更全面的总体规划来解决这个问题。总体规划不仅应包含设计方案、技术细节和成本，还需包括规划、施工和检修各阶段的时间进度表。如果拥有一份与政府商定的明确的施工进度表，这应该能增加私人投资方对投资的信任。需要特别注意的是，进度表里应该预留充足的前期现场分析和方案设计时间。当前武汉的海绵城市项目因当初仅给了3个月的设计和规划期限而深受其苦[2]。

3 未来的对策

中国现有的洪水防御相关研究建议：可以通过公众教育，促进中央政府和地方当局之间的合作，加强防汛通信预警系统以减少洪水灾害[2-3]。这些建议同样适用于引进LID和绿色屋顶削减城市雨水径流的情况。但目前绿色屋顶仍被认为是高投入，并且有可能导致水资源破坏的措施。另外，大部分国际绿色屋顶制造商常常会极力倡导绿色屋顶的雨水管理效益，中国缺乏这样成熟的绿色屋顶制造行业。进口的绿色屋顶材料运输和建造成本都很高，因此需要尽快运用本土产品构建中国绿色屋顶市场，这样不仅能够降低进口成本，还能根据城市基地特性，因地制宜地调整屋顶工程方案和用材。地方的LID持续运营需求同样需要重视，具体来说是需要海绵城市项目总体设计和建设指南。戴莉萍在武汉指出：“现有的设计建设指南只注重施工，忽略了大量后期的运营和维护。[2]”景观设计师们应该在设计阶段就做出项目运营和维护的日程表以平衡施工和运营维护之间的关系，而且强调这一长期的花费应当纳入地方当局每年的年度预算之中。LID的施工和维护应由训练有素的专业人员进行，因为绿色屋顶系统属于工程系统，需要定期检查以确保其运作达到设计目标。现有研究表明，中国当下洪水防御和LID方面的专家十分紧缺，国内外的设计教育者，制造商和LID施工安装人员可以抓住这个机会，从中央到地方开展各级培训。最后，邀请曾参与欧洲或北美地区创新且成功运营的LID项目的设计师、工程师和安装人员到中国交流想法和知识应该也会大有裨益（图6～11）。

5景天屋顶Sedum Roof5-1 细节Detail5-2 屋顶Roof

4 结论

中国近期海绵城市建设将雨水作为设计气候适应型城市时的创新设计约束条件，开创了世界先例。不论是在防洪历史、复杂的城市防洪政策发展还是在管理结构方面，中国都是世界上最古老的国家之一[5]。当下中国的海绵城市战略有利于将绿色屋顶纳入雨洪管理综合系统，作为雨洪治理的主要手段之一发挥效益，尤其是在高密度不透水的市中心区域，因为它可能是为数不多能够与公共空间灵活结合使用的雨水截流装置。绿色屋顶紧急溢流的情况始终存在，因此我们需要设计传统的雨水沟或基于LID技术的截流、滞留池来配合使用。在新城市设计阶段，地上和地下的水处理都始终需要关注[2]。

The achievements of Low Impact Development(LID) in China since the sponge city initiative was launched in 2015 in 16 cities have been extensive[1](Fig.1). The goal of LID is to emulate the stormwater/rainwater runoff function that a site had in its natural state, before alteration by humans. LID techniques include: living roofs, bio-retention areas, permeable paving, bio swales, water retention/detention ponds,rainwater harvesting and storage(Fig.2). A recent article by Dai Liping explores in detail the pros and cons of the governance structure behind the new city program in China and its financing challenges[2].The Chinese government at all levels tried to reduce flooding of its cities for centuries with structural measures such as dams and non-structural measures such as policies. Deadly floods have occurred regularly for thousands of years[3]. Weiwei Shao listed in his article on sponge cities examples of major flooding events over the past 8 years, detailing lives lost and the cost of economic damage to demonstrate how important it is to have a detailed sponge city data plan to improve city design and construction[4]. He clarified through his article that “the construction of sponge cities is a very complex task of systems engineering”[4]and acknowledged the importance of sponge city data collection to improve flood protection performance so that the sponge city initiative would be successful .

Managing floods is not new in China. In fact,for over 5000 years the Chinese have had to be extremely creative in dealing with flood protection[3,5].Wu Qingzhou’s article from 1989 explains in detail not only the history, governance structure and policies,but also China’s technical genius and willingness for experimentation in initiating flood protection in urban planning policies of its ancient cities[5]. This long history of flood protection stems partially from the geographical location of many urban settlements in the vast river deltas in Eastern China and along the coast, both regions inundated regularly with dramatic floods due in part to seasonal monsoons (Fig. 3). The precipitation pattern in China decreases from the Southeast to low in the Northwest[3].

Today additional challenges complicate matters which have primarily led to the rapid initiation of the sponge city initiative in 2015. Land use practices have changed as more and more agricultural land is used for urbanization due to China’s population growth[3].Intense construction of entire new cities populated with millions of people in the “open field”, spurred by the economic boom, has created vast areas of impervious surfaces in cities, impeding soil absorption of rain and stormwater. Additionally, climate change has made rainfall events more uncertain[3]. Shorter rainfall duration and higher intensity events increase stormwater peak flow rates resulting in uncontrollable surface- flooding and extensive water pollution of the soil, rivers and the ocean[3].

But how can these problems be actively addressed in China with such a long flood protection history? The Chinese are used to water and have lived harmoniously with rivers and lakes in their water towns such as the Jiangnan’s heritage water towns Wuzhen, Nanxun or Tongli for centuries (Fig. 4)[6]. One of the answers might lie in the hands of the public and decision makers’willingness to accept, as they did with water towns in the past, that water, in this case rain and stormwater runoff in cities, is not only a threat but a valuable resource to be protected and designed for as part of a holistic stormwater management system called LID. These LID systems slow down the stormwater to reduce increased peak flows, mitigated at source on site and recycled for reuse. It is here where “extensive” living roofs,as one of the LID tools in cities, can play a key role in partially managing rain and stormwater.Impervious downtowns especially will benefit from the stormwater mitigation effects from living roofs,as often there is not the space to manage water at grade alone due to lack of pervious surfaces.

This issue is a task where landscape architects are well positioned to address and take on a leading role in LID advocacy and visualization, and not only in its implementation. Landscape architects can design and pre-dimension LID, engineers are responsible to calculate the final dimensions of LID, and both need to work hand in hand from the outset to create a successful LID system. Landscape architects are equipped with a holistic understanding and reading of region, urban and site scales, and the complex interconnected environmental systems. They also have the knowledge and knowhow to interpret the other parameters needed to implement LID: local climate, soil drainage capacity, overall site conditions,grading and drainage, and site context. Finally, they have the know-how about which LID tools should be applied and in what combination to provide the most benefits. It could be argued if applied in the initial overall urban planning process, not as an afterthought,that LID could determine the allowable building footprint to manage stormwater on site. With China’s leading-edge sponge city plan in place and a rigorous policy applying extensive living roofs in cities as a stormwater management tool, their cities could be at the forefront in LID design and research very soon.

1 Extensive international living roof research

Research on “extensive” living roofs stormwater capabilities have been very rigorous and successful over the last years. “Extensive” living roofs are: “a single or multilayered living roof system (drainage,growing media and plants) designed individually or as composite, manufactured system. The growing media depth may be 25-150mm, while the surface may be either level or graded with low topographic features.Extensive living roofs are vegetated with typically low-growing plants”[7]. This article addresses only“extensive” living roofs as “intensive” living roofs do not qualify as economically and technically acceptable stormwater management tools for cities and are too complex to construct and maintain.

Living roofs water mitigation capacity depends mainly on local climate, evapotranspiration rate(ET), engineered soil medium, roof assembly and vegetation. Plants and growing conditions should match climate conditions to have a successfully functioning roof. Elizabeth Fassman-Beck states:“Growing medium is perhaps the most important functional component of a living roof system in terms of stormwater management - growing medium is like a sponge”[7]. Living roofs are not successful stormwater management tools alone. If the climate is extreme, for example long dry summer periods where the vegetation needs extra irrigation to survive, water has to be stored and distributed using energy. Or, extreme rainfall conditions may require additional storage in the building to detain the roof run off, if there is no space for LID at grade. Living roofs are only able to capture routine storm events. Virginia Stovin a Reader at the University of Sheffield concludes in her 2015 paper that her test plots could: “offer good retention performance for routine storm events …. none is reliably able to achieve a 2 l/s/hm2peak runoff requirement (a green field run off equivalent) for large events (i.e. measured events where P＞ 10mm or a simulated 1 in 30 year return period event)”[8].Living roofs are only one tool and should be applied with other LID tools at grade, such as bio-retention areas, swales and retention/detention ponds creating a treatment train, managing and possibly cleaning stormwater. At grade LID tools could deal with roof water peak runoff, when the extensive living roof reaches field capacity and its runoff acts like a conventional roof. In China living roofs are still not fully accepted by the design profession nor the government, as there is still hesitation about their stormwater capabilities and cost. Most research in China is initiated by individual pioneering efforts such as that of Prof. Dr.-Ing. Dong Nannan and Prof. Dr. Luo Tianqing both Tongji University who believe in their stormwater capabilities.

6 停车场改造—雨水削减实验结果Carpark retrofit—water mitigation results

In the last years significant international research has proven living roofs stormwater management capabilities. Stovin and her team,for example, have been testing the hydrological performance of soil media (substrate) and vegetation configuration in test beds at the University of Sheffield for years and conclude in one of their latest papers 2015: “Overall the study has demonstrated that the configurations most typical of commercial extensive green roof systems,i.e. Sedum vegetation on a brick-based substrate,will offer the best all-round performance in terms of both retention and detention. However, it should be noted that shallow, extensive, green roof systems need to be combined with downstream retention and detention measures to provide more holistic Sustainable Drainage Systems (SuDS)solutions that can mitigate flood risk for even the largest storm events”[8]. She also observed: “that long term vegetated living roofs with well graded,less permeable substrate have a higher retention capacity than large-pored permeable substrate”[8].

In Germany Sedum has been seen as the “work horse” for living roofs successfully for 40 years (Fig.5)[9]. This new research confirms again that this species in combination with less permeable substrate is useful for stormwater retention. Another team in Australia reviewed how plant composition and water use strategies could in fluence stormwater retention and concluded: “Contrary to our hypotheses, green roof modules planted with species with water use plasticity did not always retain more water than the conservative water using succulent,Sedum pachyphyllum. Also, the mixtures of plant species did not retain more water than monocultures. Although the mixed species treatment had higher biomass and productivity, the performance of the mixture was adversely affected by the key species,Stypandra glauca. The root systems ofS. glaucalikely created preferential flow pathways through the substrate,limiting substrate water contents, evapotranspiration,and rainfall retention. To optimize rainfall retention and plant survival, species selection for green roofs should therefore not only consider water use strategies or species diversity. Root traits should also be considered when selecting green roof plants, due to the importance of preferential flows overriding the benefits of plastic plant water use strategies”[10].

7 停车场改造屋顶平面图—设计阶段Carpark retrofit roof plan—programming

8 停车场改造场地平面图—LID系统布局Carpark retrofit site plan—LID systems layout

9 轴测图Axon

These two current examples demonstrate the depth, detail and rigor of existing research on living roof water retention capacities and its limitations. It is here where the opportunity for China lies with the sponge city initiative. Most of the experiments like the two above have been done in controlled smallscale test bed environments on roofs and at grade.China has now the opportunity with the sponge city initiative to expand these experiments to an urban scale including monitoring the constructed living roofs from the outset. This would not only continue their historic leading-edge tradition of flood control research but it might help to safeguard the cities from major floods and entice the public to see rain and stormwater as a positive element in urban design. Living roofs are at the front end of the LID treatment train. As mentioned above, alone they are not a reliable tool for large storm events. Therefore,it is important to design and construct living roofs at urban scale in combination with at grade LID.

Living roof benefits for LID include:1) Reducing rainwater runoff entering public stormwater sewers through plants and soils evapotranspiration. 2) Reducing upgrades of existing public stormwater sewers needed otherwise due to climate change. 3) Cleaning runoff and recharging the water table at source in dense urban areas.4) Providing a continuous ecological habitat from built structures to grade and vice versa. 5) Providing a “viewing experience”, that can also have a positive impact on health. 6) Improving the microclimate around adjacent buildings.

2 Designer responsibility

Landscape architects are potentially liable when designing living roofs and LID at grade. It’s a serious matter, if it goes wrong, and the economic and reputation damage to an of fice and the profession could be significant. Therefore, a rigorous understanding is needed to design and supervise their construction. This is indeed common knowledge but landscape architects could do more earlier on in the design process. They should be involved in living roof advocacy, and at the initial urban planning stage of any design project. In fact, they should be the first designers on site together with urban designers and architects. The sponge cities concept provides an excellent opportunity for landscape architects to take a leading role. LID should not be an afterthought; in fact, it should be the first question when a project is initiated. Liping Dai describes the governance structure of the sponge city program in his article from the top-down, from central government to local government[2]. He emphasizes that preparing cities for flooding is not only a technical but also- especially at the beginning of planning a sponge city—a governance issue[2]. Designers should try to recognize where they are working and understand these structures to provide LID solutions acceptable to these governance structures. Ling Dai divides China’s governance into four categories: governing by example,authority, provision, and through (co)funding[2]. This structure could provide a framework for where and how the designers could be involved in all four.

10 停车场改造屋顶LID系统—细节和装置Carpark retrofit roof LID system—details &assembly

11-1 露天剧场南北剖面North-south section at amphitheatre11-2 啤酒花园南北剖面Noth-south section at Beer Garden

Governing by example could be the sponge city initiative in itself, and analyzing already built projects such as Liping Dai’s paper does with the Wuhan case study[2]. Landscape architects could provide LID visualizations emphasizing living roofs stormwater benefits, explanatory animated online movies explaining the holistic LID systems,public talks in TED talk manner to elevate the importance of LID and starting a new culture of public presentations in local municipalities including the residents. Currently public participation is not compulsory in sponge city policies in China[2]. This has to do with traditional Chinese culture were the national interest is above the collective and ‘lastly’individual interest. These principals are rooted in the Chinese planning ideology and will need to be reviewed closely in the future when designing LID[2],as at the moment national interests still prevail.Landscape architects should not be deterred from promoting their LID skills and knowledge and seek opportunities to integrate public participation through informing them through publications,discussions, conferences, public talks, exhibitions and Chinese social media. Designers should be public advocates of their skills and ideas. In governing by authority where traditional regulations and sanctions are used[2], landscape architects could advocate to help shape those regulations including LID tools as a necessity, and emphasize living roofs stormwater management benefits, especially in dense city areas.Existing international and recent national peerreviewed research could help make a convincing argument. Technical guidelines should be tailored to local geographic conditions[2]. In governing by provision, designers could provide LID resources,toolkits such as LID calculators and international technical knowledge to authorities. “Education” in reading project sites, existing stormwater conditions and their context at the beginning of the planning phase is important for LID decisions made by clients, governmental employees, public and future installers of LID. Landscape architects are the link between researchers on one side and clients,authorities and installers on the other. Governance through (co)funding is difficult in China as private investors distrust the government[2]. In this instance landscape architects could provide a masterplan which not only includes design, technical details and cost, but an overview plan for the planning,construction and maintenance schedules. This might reduce the current distrust by private investors to invest money, if a clear construction timeline is agreed upon with the government. Important to note is to include ample preliminary site analysis and design time, as current sponge city projects in Wuhan suffered from very short 3-month design and planning deadlines[2].

3 Future steps

Current research on flood control and prevention in China recommends: citizen education, promoting cooperation between federal governments and local authorities, and functioning flood communication systems to reduce flood damage[2-3]. These recommendations could also be applied to introduce LID and living roofs in cities to reduce stormwater runoff. But at the moment living roofs are still seen as expensive and a risk for water damage. Additionally,there is no established living roof industry in China to advocate their benefits as mostly international manufacturers are used. Material transportation and construction costs for those roofs are high and it would make sense to initiate a Chinese living roof market using local products to reduce import costs and adapt the engineered roof assembly to local conditions and availability of materials. Also of importance are LID’s local ongoing operational needs. Specifically, for sponge projects’ general design and construction guidelines,Liping Dia noted in Wuhan: “that they ignored largely operation and maintenance, while the focus was mainly on their construction”[2]. Landscape architects could provide project operation and maintenance schedules as part of their design package to balance out the current focus on construction and emphasize that this longterm cost should be included in the local authorities’annual budget. LID construction and maintenance should be carried out by trained staff, as living roofs and at grade systems, being engineered systems, need regular checks to make sure they function to their design objectives. Current research notes there is a lack of trained flood control professionals including LID specialists in China, which is an opportunity for international and local design educators, manufactures and installers to provide training from federal to local levels. Finally, it would be useful to invite the designers,engineers and installers involved with innovative wellfunctioning LID projects in Europe or North America,to China, to exchange ideas and knowledge (Fig. 6-11).

4 Conclusion

China’s recent sponge city initiative sets a precedent for the rest of the world to use storm and rainwater as a creative design constraint when designing cities for climate change adaption. China’s flood prevention history, sophisticated urban flood policy development and governance structure is one of the oldest[5]. The current sponge projects are a great opportunity now to integrate living roofs stormwater management benefits as one of the main tools of LID,especially in dense impervious downtowns, where they may be one of the only tools in conjunction with public spaces used as flexible water retention facilities.Living roof runoff should always have an emergency overflow, we therefore are going to need to continue designing conventional storm drains or retention/detention ponds in conjunction with LID. We should always focus on both above ground and underground services when we design new cities[2].

Notes:

① This paper is revised on the basis of the speech delivered by the author at the International Landscape Architecture Symposium in 2017.

② Fig. 1 is quoted from https://cleantechnica.com; Fig. 2 ©reference [7]; Fig. 2 is quoted from https://www.express.co.uk;Fig. 4 is quoted from https://zx.sztravel.gov.cn; Fig. 5 © Daniel Roehr; Fig. 6-11: Living roofs and LID should be integrated in all urban design, architectural, landscape architectural and planning syllabi. Drawings from the University of British Columbia, Vancouver, Canada “Advanced Living Roof Seminar LARC 582E” final UBC SEEDS Report 2018. Drawn by Jericho Bankston, Josh Harvey, Instructor: Assoc. Prof.Daniel Roehr.

(Editor / WANG Yilan)

注释：

① 本文以作者在2017世界风景园林师高峰讲坛上的发言稿为基础进行补充。

② 图1来自https://cleantechnica.com；图2引自参考文献[7]；图3来自https://www.express.co.uk；图4来自https://zx.sztravel.gov.cn；图5由作者拍摄；图6～11：绿色屋顶和低影响设计，来自2018英属哥伦比亚大学（UBC）社会生态经济发展研究（SEEDS）LARC 582E报告，由杰里科·班斯顿、乔什·哈维绘制，指导老师为丹尼尔·勒尔副教授。

（编辑/王一兰）