Urbanization has been the dominant trend of societal development all over the World during the last half century. Accompanying the economic growth and the accelerating urbanization, the threat of water issues such as water shortages, lack of the access to clean and safe water, etc., has become one of the major constraints on the development in many cities and regions [1,2,3,4]. Water shortages are reflected in not only the water quantity but also the water quality. However, the combined effects of point source (PS) and nonpoint source (NPS) pollution of the water environment make it difficult to fundamentally improve water quality [5,6]. Because of the increasing enforcement of PS pollution control and the difficulty to specifically identify or abate the NPS pollution, the proportion of NPS pollution in water contamination obviously increases and contributes to a large fraction of the water impairment.

In the majority of regions in the World, urban NPS pollution is the second important component of NPS pollution following agricultural NPS pollution [7]. In the USA, for instance, water quality in 13 percent of the rivers and streams, 18 percent of the lakes, ponds, and reservoirs, and 32 percent of the estuaries was considered impaired, mainly by urban NPS pollution [8].

The accelerating urbanization and the resulting increasing impervious surfaces are supposed to be the dominant driver for urban NPS pollution. Impervious surfaces prevent water from seeping into the ground and cause the occurrence of surface runoff. With the process of urbanization, the amount of impervious surfaces is increasing rapidly, thus more runoff is created and less water is able to sink in, or “infiltrate” into the ground. Water that travels slowly through the ground gets filtered by natural processes before it reaches the receiving water body, but water that travels too quickly to creeks and streams can pick up and carry a lot more sediment and other pollutants.

Usually, intense rainfalls are assumed to be the major flood-generating events in urban areas. However, flooding is also observed during snowmelt in urban environments in Scandinavia, Canada and the USA [9,10,11]. Flooding arising from snowmelts in early spring can also be observed in Northern China. Nevertheless, in contrast to the rainfall, snow is not mentioned in many urban NPS pollution study cases. That is probably because melt intensities are much lower than rain intensities, and because processes controlling snowmelt are purported to be the same in urban and rural basins. However, for snow, there are different processes and factors controlling runoff. Also, the snowmelt-related runoff conditions in a city are different in winter compared to storm runoff in summer [9]. On the other hand, the extreme weather associated with the global warming may lead more precipitation in some cold regions [12]. In other regions, the increasing temperatures may affect the length and time of the snow season and snow cover duration [13]. Accordingly, the load and characteristic of the snowmelt-related runoff may change obviously, which make it a major urban nonpoint pollution source and carrier.

Therefore, the aim of this study was to review and examine the urban NPS pollution problem considering both the rainfall and the snowmelt runoff, highlight the importance of the snowmelt runoff in urban NPS pollution against the background of global climate change, and to broach the foreground and development tendency of the snowmelt runoff research and management. The current work is important for both protecting and economizing the urban water resources and completing the theory of urban NPS pollution in cold regions.

The research of urban NPS pollution began in the 1960s–1970s in developed countries, and subsequently a set of models and management systems based on the rainfall runoff characteristics and pollution load emerged. Storm water runoff discharge laws have been established in the USA, which requested that measures be taken during certain exploitation actions to control surface runoff, such as receiving water after storms being qualified [14]. In 2005, the management of urban NPS pollution was highlighted and the status of urban NPS pollution was changed from a local problem to a national problem in the US, marked by the decreeing of the US EPA government National Pollutant Discharge Elimination System (NPDES) program, which requires municipalities of a certain size to apply for permits to discharge storm water. The supporting documentation contains some assessment of the urban runoff pollution, including the results of field sampling, and proposes control measures. However, it was not until early the 1980s that urban NPS pollution was taken into consideration in China. The research work was firstly carried out in Beijing, and then in Shanghai, Hangzhou, Suzhou, etc. [15,16]. Limited cases can be found in Northeast China, except for a case study in Siping City [17].

Urban NPS pollution shares all the common characteristics of NPS pollution, however, obvious variations can be found between urban NPS pollution and the agricultural NPS pollution because of the difference in landform and surface features, land use patterns and anthropogenic activities [15]. The major characteristics of urban NPS pollution are reflected in the randomness of occurrence, the indefinite and extensive distribution, the complex processes, the hysteresis, the “first flush” effect for rainfall runoff (80 percent of the pollution volume is carried by 30 percent of the first runoff volume) [15,18], and the “first melt” effect for snowmelt runoff (the first action results in immediate runoff, usually involving small volumes of water and a minor portion of the annual pollution load, although concentrations may be high) [19], etc. For urban NPS pollution, the major pollutants include sediments, nutrients, heavy metals, oils and hydrocarbons, and oxygen-demanding substances, etc. [16]. Many factors can affect the characteristics and the load of urban NPS pollution, in which, the climate and hydrology, air quality, preceeding dry weather period, land use patterns, area of the impervious surfaces, urban drainage systems and urban sanitation all feature largely among the factors [20,21].

Although most of the studies pertaining to urban NPS pollution were focused on the runoff arising from big rainstorm events, there was tangible evidence that snowmelt runoff was responsible for a big fraction of urban environment contamination and the receiving water impairment. Thorolfsson and Brandt [11] studied urban storm runoff during summer and winter in Norway from 1988 through 1994 and found that snowmelt runoff had a much greater in volume than typically considered in drainage designs, resulting in much more winter flooding than that in summer. Besides, an urban storm-runoff model that considers snowmelt and rainfall was also produced [11], but it was concluded that there was still a notable lack of experience about urban storm runoff during the winter season. On the other hand, despite previous conclusions that not all kinds of pollutants were prone to dissolve in snowmelt [22], snow has been shown to store high levels of various pollutants including lead, hydrocarbons, polychlorinated biphenyls, sulfate and other metals and solids [23,24]. For instance, Blecken et al. [25] found that due to the change of runoff intensity and high sediment loads in the snowmelt runoff, snowmelt runoff can cause an increased proportion of fine-grained sediment fractions (<0.063 mm) in spring, and the retention of metals can also be observed because of the low turbulence in the water and the presence of organic material,. Thus the snowmelt runoff indeed must be taken into consideration in studying and controlling the urban NPS pollution. However, it is difficult to identify and control the urban snowmelt runoff due to the anthropogenic redistribution activities, the complex melting process, the variety of runoff condition in cities and the global climate change.

Models play critical roles in urban NPS pollution research. On-site measurement and monitoring of urban NPS pollution is expensive and time consuming. Runoff events are intermittent, and long-term records would be required to measure the runoff from a specific site. In addition, there are many complicating factors affecting urban NPS pollution. For these reasons, models are, in most cases, the only reasonable tools for runoff prediction, assessment, and management. Earlier models for urban NPS pollution were mostly performed based on the river quality change arising from land utilization. Most of them belong to the statistic models class. The models mainly focus on the rainfall runoff characteristics, individual rainstorms and long-term runoff pollution load, and the relationship between runoff amount and the land utilization. The most famous models include the Storm Water Management Model (SWMM), the urban surface runoff model (STORM), the Wallingford model and the MOUSE model.

The SWMM was first developed back in 1971 by the US EPA and has undergone several major upgrades. The model is a dynamic rainfall-runoff simulation model used for single event or long-term simulation of runoff quantity and quality from primarily urban areas. It continues to be widely used throughout the World for planning, analysis and design related to storm water runoff, combined sewers, sanitary sewers, and other drainage systems in urban areas, with many applications in non-urban areas as well [26]. The STORM model, developed in 1973, is designed to model urban watersheds and is capable of calculating loads and concentrations of water quality parameters, such as suspended and settleable solids, biochemical oxygen demand, total nitrogen, orthophosphate, and total coliforms, besides, it is also capable of calculating land surface erosion. STORM requires hourly precipitation data to model the stormwater elements which result in a difficulty of collecting the vast basic data needed. The Wallingford Model was explored in 1978 by the Institute of Hydraulics, Wallingford, U.K. The model can simulate a storm and waste water system and was widely used in the operation, design and layout of storm equipment. The MOUSE model developed by DHI was a synthetic model which can simulate the hydrology, the hydraulic and the water quality of urban sewerage and drainage systems [27]. Besides the above classical models, many new models based on rich experience appeared in the 1980s–1990s. The new models tend to focus on the management and risk assessment of NPS pollution, which greatly promoted urban NPS pollution model research. With the development of computer and 3S techniques, models with powerful functions, such as special data management, database technique and digital expression were explored based on drainage scale, providing advantages to urban NPS pollution research.

The importance of snowmelt runoff in melting seasons emerged gradually in the recent decades, especially in cold regions, thus some urban NPS pollution models were explored specially for the snowmelt runoff events. One of the most typical models is the Urban Snow Model (USM) built by Valeo and Ho in 2004 [28]. USM uses an energy balance scheme at an hourly time step, changes in urban snow albedo, and incorporates eight different types of redistributed snow cover. The full energy budget version of USM outperformed all other models in terms of time to peak, peak flow rate and model efficiency. Its modified version is recommended when a lack of data exists.

Conceptual snowmelt runoff models have proven useful on estimating discharge from remote mountain areas, including the various spanning ranges of the Himalayas. Such models can provide fairly accurate predictions of water availability for operational purposes (e.g., irrigation and hydropower). However, the capability of these models to address characteristic components of water disputes such as diversions, storage and withholding is limited. Contemporary disputes between India and Pakistan are partly due to the snowmelt-derived water resources from the Upper Indus Basin, and water balance accounting methods need to be improved to cope with this issue. Kult et al. [29] presented a research agenda focused on providing refined hydrological contributions to solve water disputes. The snowmelt runoff model built by Tahir et al. [30] in 2011 can be used efficiently in the snow- and glacier-fed sub-catchments of the Upper Indus River Basin (UIB). The application of the snowmelt runoff model under future climate (mean temperature, precipitation and snow cover) change scenarios indicated a doubling of summer runoff until the middle of this century.

However, for most countries and regions, especially the countries where urban NPS pollution research is in an infancy status, the models fail to be applied suitably. For instance, most of the models concerning surface runoff have been widely applied in China during the last few years, but the urban development mode and basic conditions in China vary largely from those of cities in other countries. The comparatively high building mass and fewer greenbelt zones result in more impervious floor area in the cities of China. On the other hand, the poor urban sanitation conditions, and the lack of basic data lead to the failure to use the models suitably. To date, there is neither a good model nor perfect experience on management and control of urban surface runoff in China [14], let alone models or methods capable of comprehensively considering both the rainstorms in summer and the snowmelt runoff in winter and early spring.

The general method of controlling urban NPS is to apply best management practices (BMPs) which has been used by many projects in many countries [46]. BMPs was defined by US EPA as any methods, practices and operation programs of mitigation and precaution in water resources pollution including the operation and maintenance program of engineering and non-engineering practices [47]. In practical application, it is adopted rationally according to local weather and natural geography conditions in different areas [14]. Besides, developed countries have built many systemic management methods for urban NPS pollution. The US EPA [48] documented a comprehensive runoff management program (Figure 2) to provide technical assistance to state and local program managers and other practitioners on the best available, most economically achievable means of managing urban runoff and reducing NPS pollution of surface and ground waters from urban sources. In most developing countries, however, the management and control of urban NPS pollution is not optimistic due to the long-term neglect of urban NPS pollution issues. In China, for instance, there is neither a policy with respect to urban NPS pollution nor laws pertaining to control of urban NPS pollution. This is not conducive to control of urban NPS pollution. Thus, the policy-making may become one of the most urgent and necessary works for controlling urban NPS pollution in developing countries. The contents of policies should include the management, precaution, the monitoring, and the rewards and punishment systems. In view of the potential risk of snowmelt runoff under the climate change, the policy should also consider both the summer flooding and the snowmelt runoff in cold regions. Local governments at all levels should also work out local policy guidance on the storm and sewage sewers design in accordance with actual conditions.

In addition to the management system, exploring the ecological purification technique is recommended for tackling urban NPS pollutants. The constructed wetlands and grass greenbelts are the preferred tools of ecological purification. Constructed wetlands have gained international interest and applications due to their low maintenance and operational costs, and their high removal capacity for many kinds of pollutants [49,50]. For instance, in the U.K., a significant amount of money has been spent to investigate reed bed schemes in England and Wales [51] to purify the water. In Sweetwater, USA, a constructed wetland was applied for treating urban runoff and it does reduce the amount of nutrients and metals in the water column [51]. In China, Liu et al. [52,53] obtained good purification ability for the NPS pollutant nitrogen with constructed wetlands. Zhang [36] used a grass swale system for purifying urban storm runoff with satisfactory results. In New Delhi, India, constructed wetlands are also applied for urban runoff control [54]. However, in general, the application of ecological purification technique in urban NPS pollution control in developing countries has lagged behind that in developed countries due to the lack of economic support and the needed technical expertise. More cases of practical application are needed, but not always stay on the experimental scale. In addition, in view of the considerable snowmelt runoff, the lower microorganism activity under cold conditions and the resulting lower purification capacity, exploring zoology purification techniques which have the capacity to be operated under cold environments is important and of far-reaching significance in this field.

In the current review study, progress of NPS pollution studies considering both the summer flooding and the snowmelt runoff was examined. The snowmelt runoff was highlighted for its more and more important status in urban NPS pollution under the conditions of global climate change. With the global temperature increase, the considerable snowmelt quantity and the worrisome water quality in cold regions make it urgent to take measures to manage snowmelt runoff in melting seasons despite of the lack of data in this field now. Policy-making and application of ecological treatment techniques are recommended. The next step of the study is to apply the runoff model to simulate and forecast the summer stream flow as well as to evaluate the impact of future climate scenarios.