### 1. Introduction

### 2. Methodology

### 2.1. Study area

### 2.2. System description

^{2}; a rainwater storage tank; a rainwater treatment units which encompass a sediment tank to deposit large solids (e.g. leaves and twigs) and a gravity driven micro-filter (GDM) system to removal suspend solids and dissolved pollutants; a clean water storage tank, and a convey system through which the treated rainwater is sent into the end user of building. After GDM process, the quality of treated rainwater can well meet the standard of Chinese Recycling Water Guidelines, which has been confirmed in our previous researches (published online). The harvested rainwater is mainly used for non-potable purposes (toilet flushing and hand washing) in the target buildings.

### 2.3. Data

#### 2.3.1. Rainfall data

#### 2.3.2. Water consumption

^{3}, corresponding to an average daily water consumption of 19 m

^{3}, This water mainly used for toilet flushing, hand washing and floor cleaning, which do not require potable water quality since there are only four kinds of sanitary appliances in the toilet of this multifunctional building, i.e. toilet, urinal, wash basin and sink. The water consumption data were then used to estimate the water saving efficiency and reduction of tap water, as well as the benefits costs analysis of the RWH system. The daily water consumption of the target building was summarized in Supplementary material.

#### 2.3.3. Economic data

^{3}(2020 Guangzhou Water price). It should be noted that the water price varies from one city to another, and therefore the outcomes of this study need to be interpreted in relation to the water price at other locations of interest.

### 2.4. Data analysis

_{c}is the volume of available rainwater (m

^{3}), R is the local daily rainfall (m), A

_{c}is the catchment area of roof (m

^{2}), R

_{c}is the surface runoff coefficient, assumed equal to 0.8 to represent losses of 20%.

^{3}).

_{t}is the volume of water saved over a period of time t (m

^{3}), Pt is the cost of water over a period of time t (CNY/m

^{3}), I

_{t}is the investment required for a period of time t (CNY), M

_{t}is the maintenance costs over a period of time t (CNY), s is the system life span (year), t is the system operation period (year), and r is the discount rate (%).

### 3. Results and analysis

### 3.1. Water savings analysis

^{3}potable water can be saved by a 1 m

^{3}storage tank, nevertheless, this is boosted to over 3,300 m

^{3}if a 15 m

^{3}storage tank is used (Fig. 6). These patterns are in agreement with previous studies [13-16]. For the first stage, it is evident that increasing the tank volume affords an opportunity to store more rainwater. However, the amount of water saving reaches a plateau since the rainwater tank is able to store the total rainwater captured from the rooftop. Therefore, it is useless to increase the tank size once it exceeding 20 m

^{3}.

^{3}, the WSE increase up to 57%. Afterwards, this indicator reaches a plateau for the range of 20 m

^{3}to 30 m

^{3}.

^{3}and a WSE of 56.86%, respectively, with a storage capacity equal to or greater than 20 m

^{3}(Fig. 6).

### 3.2. Payback period analysis

^{3}to 10 m

^{3}, then increases when tank size ranges from 10 m

^{3}to 30 m

^{3}, similar to the indicator of water saving obtained within these ranges of storage tanks, a positive relationship between the payback period of investment and tank capacity was observed, i.e., the greater the tank size, the shorter the PBP could achieved. The PBP values of RWH system for 1 m

^{3}and 2 m

^{3}schemes are incomputable according to formula (3). As illustrates in Fig. 7, the payback period for 3 m

^{3}tank is more than 50 years. Whereas it is commonly to regard an investments project as economically viable when the payback periods of which are less than its lifespan. For office buildings, the service life is generally within 40 years in China, therefore, the RWH system with a storage tank less than or equal to 3 m

^{3}are consider to be economically unfeasible. As Fig. 7 illustrates, the payback period of RWH system can be shortened to lower than 18 years if the tank capacity increases to 4 m

^{3}. The PBP values are all shorter than the threshold of 40 years with rainwater tank range from 4 m

^{3}to 30 m

^{3}, while a RWH system within the range of 6 m

^{3}to 25 m

^{3}tank even gives a PBP of lower than 10 years. Obviously, the shortest PBP (6.26 years) is achieved for a 10 m

^{3}rainwater tank scheme.

^{3}, this is due to the fact that increases tank size raises initial investment and operating costs of the RWH system, while the amount of captured rainwater does not increase with the increase of tank capacity since the available rainwater is limited. Thus a successive increase in tank size will eventually leading to a point that the annual increase in expenditures exceeds the increase in revenues.

^{3}, for instance, a nearly 80% decrease in payback period was found when rainwater capacity increased from 3 m

^{3}to 5 m

^{3}. Nevertheless, the payback period decreased by only 26.9% when rainwater capacity increased from 6 m

^{3}to 8 m

^{3}, and only a 5.6% increase in PBP was witnessed with tank size increasing from 8 m

^{3}to 20 m

^{3}. This is because, with the continuous increasing of storage capacity, the tank size became a less limited factor, instead, the rainfall will becomes a more relevant variable for benefit effectiveness of an RWH system.

### 3.3. Benefit cost ratios analysis

^{3}, then the BCR reaches a plateau of around 1.50 when the tank sizes is between 10 m

^{3}and 15 m

^{3}. Afterwards, the BCR values start to decrease from 1.50 when the tank size is range from 15 m

^{3}to 30 m

^{3}, which shows a similar pattern as the PBP does. The current work found that the maximum BCR (1.50) of the RWH system is achieved for a 10 m

^{3}tank scheme. The present study also found that the benefit cost ratio were below the desired value of 1.00 for small tank scenarios (between 1 m

^{3}and 4 m

^{3}).

### 3.4. Sensitivity analysis

^{3}tank installed in the target building. The water price considered in this study is 3.88 CNY/m

^{3}(2019 Guangzhou water price) and both positive and negative variations of the water prices were considered.

^{3}(2020 Guangzhou water price). Table 3 illustrates that a 10% increase in the current water price (from 3.88 to 4.27 CNY/m

^{3}) would give a benefit cost ratio of 3.69 (which represents a 10.15% increase) for the RWH system, while a 20% increase in the current water price (from 3.88 to 4.66 CNY/m

^{3}) would give a benefit cost ratio of 4.02 (which represents a 20% increase), presenting a clear linear relationship between the water price and BCR values. However, an variation of -20% to +20% in the initial investment resulted in variations of the benefits costs ratio from +5% to -4.5%, and an 20% increase in the annual operating costs (from 2,655.64 CNY per year to 3,186.77 CNY/ per year) only leading to a 10.4% decrease in BCR for this RWH system (Table 3). These results demonstrates that water price is the most important factor affecting the return on investment, followed by operating costs and initial investments.

### 4. Discussion

^{3}. The economic indicators in term of benefit cost ratio is 1.50 and payback periods are within 6.26 years, respectively, implying that this green infrastructures is economically viable for the a public building in Guangzhou.

^{3}), the tank capacity is a limiting factor. It is evident that increasing the tank capacity within these ranges offered an opportunity to store more rainwater. However, with a further increase in tank sizes, the annual water saving and WSE reached a plateau of 3,923.56 m

^{3}and 56.86%, respectively. In this case, tank capacity is no longer a limiting factor. Instead, the catchment area of target building becomes a new constraint to the water saving potential of the RWH system, Hence, it will be uneconomical to increase the tank capacity considering the total amount of rainwater available is limited.

^{3}tank scheme gives a PBP of 56.4 years. The high value of PBP (more than 40 years) implying that the RWH system is not economically viable in that case.

^{3}, after that, the BCR values reaches a plateau between 1.45 and 1.49 for tank sizes range from 8 m

^{3}to 15 m

^{3}. Beyond this, the BCR values start to decrease when the tank size is increased from 15 m

^{3}to 30 m

^{3}. but still show positive values (see Fig. 8), implying that further investment for increasing tank capacity gives less benefit compared to the cost.

^{3}tank. Whereas in another study, Tam et al. (2010) [19] investigated the cost effectiveness of RWH system in residential houses around Australia and found that although this system can offer a considerable portion of non-drinking water, but is hard to achieve desirable financial benefit for Brisbane. The same study found that it could not be possible to achieve “pay back” for the RWH system without some favourable scenarios and conditions for multi-storey buildings in Sydney [9]. Literature also reported that rainwater collection would only be feasible in South Korea for during 6 months of the year [20]. As has been noted, the financial viability of an RWH system depending on the tank size, climatic conditions, non-potable water consumption pattern, public water price as well as a discount rate considered. Our study found that for a multifunctional building in Guangzhou, the total investment of an RWH system can be recovered within 6.26-16.56 years time depending on the tank size. The humid climates in South China (with average annual precipitation higher than 1,700 mm) is the main reason account for these high costs effectiveness of an RWH system in Guangzhou. This finding is also consistent with the results found by Imteaz [21] and Jing [22], as they stated that the local rainfall is the major variables of interest in the design and evaluation of RWH systems, as it determines the amount of collectible rainwater generated from a given contributing catchment.