食物是保障居民日常生活的基本消费品，它的消费伴随着大量的水资源消耗。近年来，随着经济发展和人民生活水平的提高，中国居民的膳食结构和膳食水足迹发生了显著的变化。探究中国居民膳食水足迹变化的驱动机制具有重要意义。当前膳食水足迹的研究内容大多为针对某个研究区进行水足迹核算或单独分析膳食水足迹与某种影响因素之间的相关性，对每种驱动因素的影响缺乏具体的量化。为深入探究引起中国居民膳食水足迹变化的驱动机制，本研究同时考虑多种因素的作用，通过因素分解来量化每种因素的贡献量，分析膳食水足迹变化的原因。考虑到膳食消费对水资源的影响，本研究基于水足迹视角，对居民膳食结构进行调整，提出有助于缓解水资源压力的膳食结构方案。 本研究首先从水足迹角度出发，对中国居民2001-2016年的膳食水足迹进行了核算，并对比分析各省份居民人均膳食水足迹的空间分布差异及年际变化；之后采用LMDI分解方法，建立膳食水足迹总量变化的因素分解模型，剖析各个驱动因素对城镇和农村之间及不同省份膳食水足迹的影响；最后在考虑人体营养需求的情况下，构建多种膳食结构情景，对各省份的居民人均膳食水足迹进行情景分析，对面向水资源压力缓解的膳食结构方案设计具有借鉴意义。 膳食水足迹核算结果表明，城镇和农村居民人均膳食水足迹（CWF）总体上均由东西两侧向中部递减，呈现“东西高-中部低”的空间分布。其中，人均膳食绿水足迹比例较高，大体呈南高北低的分布特征；人均膳食蓝水足迹的较高值，大多出现在华北、西北和东部地区；人均膳食灰水足迹，总体上空间差异不大。就城镇和农村差异而言，在研究期内各省的城镇居民人均膳食水足迹普遍高于农村居民。2001-2016年间，中国居民膳食水足迹总量（TWF）由5534.22亿立方米增加到6612.86亿立方米，并且近年来增长趋势更加明显，说明中国膳食消费尚有较大的节水潜力。就膳食水足迹的组成结构而言，粮食消费水足迹在TWF中的比例，由2001年的47.4%下降到2016年的31.4%，而肉禽类的消费水足迹比例从23.6%上升至32.24%。 基于上述核算结果，本研究对2001-2016年TWF的变化量进行了驱动因素分析，结果表明经济水平是膳食水足迹增加的最大正向因素，其贡献量为8.2×1011 m3，相当于2001年中国居民膳食水足迹总量的1.48倍。用水效率是一个重要的抑制因素，使膳食水足迹降低7.69×1011 m3，说明技术水平和用水效率的提高为减少水资源消耗发挥了重要作用。考虑到膳食消费对水资源的影响，本研究基于水足迹视角，兼顾人体营养需求，构建多种膳食结构情景，对膳食水足迹进行情景分析。结果发现由豆类食物代替肉类、水产品，在保证人体能量、蛋白质等营养需求的同时，使居民人均膳食水足迹明显下降。由此表明选择多植物性、少动物性食物的膳食结构，对于从需求侧降低水资源消耗具有重要意义。

Food is a basic consumer product that protects the daily lives of residents; its consumption causes a large amount of water consumption. In recent years, the Chinese inhabitants’ dietary structure and dietary water footprint have changed significantly. Most of the current study on dietary water footprint is to calculate the water footprint of a certain area or to analyze the correlation between the dietary water footprint and a factor, lacking quantitative study on the impact of several drivers. In order to explore the driving mechanism which caused the change of dietary water footprint of Chinese residents, the effects of multiple factors are considered and the contribution of each factor are quantifier by driving force decomposition. To alleviate the pressure of water resources, the dietary structure of residents was adjusted from the water footprint perspective. First, this study calculates the dietary water footprint of Chinese residents during 2001-2016; second, the temporal and spatial changes of the dietary water footprint are analyzed. And then LMDI method is used to explore the driving mechanism of changed dietary water footprint in different years. Finally, three dietary structure scenarios in consideration of human nutrition needs are established to evaluate the per capita dietary water footprint. It can be used as a reference for the design of dietary structure for the water resources pressure relief. The results show that the per capita dietary water footprint (CWF) of urban and rural inhabitants decreased from both sides of the east and west to the middle in general. The spatial distribution characteristics of per capita green water footprint are similar to those of CWF; most of the higher values of dietary blue water footprint per capita occurred in North, Northwest and Eastern China. As far as urban and rural differences are concerned, the per capita dietary water footprint of urban residents in most provinces is higher than that of rural. During 2001-2016, the total dietary water footprint (TWF) of Chinese residents increased from 5.53×1011 m3 to 6.61×1011 m3, and the growth trend has become more apparent in recent years, indicating that there is still a great potential for saving water in Chinese dietary consumption. In terms of the structure of dietary water footprint, the proportion of water footprint of grain consumption in TWF has decreased from 47.4% in 2001 to 31.4% in 2016, while the proportion of water footprint of meat and poultry consumption has increased from 23.6% to 32.24%. The analyses of drivers of total dietary water footprint change illustrate that the economic level is the largest positive factor in the increase of dietary water footprint. Its contribution is 1.48 times as big as the total amount of dietary water footprint of Chinese residents in 2001. The water use efficiency is a negative factor which leads to a decrease of 7.69×1011 m3. It indicates that the improvement of the water use efficiency has played an important role in reducing water consumption. The results of scenario analysis of dietary structure indicate that dietary water footprint of residents will decrease when meat and aquatic products are substituted by bean. Therefore, it is meaningful for water saving from the demand side to choose the diet with more plants and less animal food.