Introduction

The temporal and spatial variability of rainfall are principal factors in all aspects of water resources management including the allocation of water supplies, management of agricultural irrigation, storm water and reservoirs, and permitting of large withdrawals [1,2,3,4]. Some outdoor water conservation policies implicitly respond to rainfall by implementing restrictions on lawn irrigation only during drought, such as when reservoir or stream stages drop below a certain point; however, outdoor water restrictions (OWR), a common lawn water conservation strategy, are increasingly employed as long-standing policies and thus neither implicitly or explicitly address rainfall. We are unaware of any OWR policies that are explicitly designed around historical trends in lawn water demand.

The typical irrigation schedule associated with OWR is based on even/odd addresses being assigned to certain days of the week when lawn watering is allowed. These weekly watering schedules, which are a central feature of OWR, provide a structured framework to communicate limits on the timing and frequency of lawn watering to end-users. The central assumption behind OWR is that a weekly irrigation schedule will result in constrained lawn watering and thereby promote water savings [5,6,7]. However, depending on weekly rainfall, even a minimal irrigation schedule can allow overwatering. The connection between rainfall, net irrigation need, and lawn watering is critical; otherwise, weekly lawn irrigation can persist even when abundant rainfall deems it unnecessary. Thus, there is a significant opportunity to improve the effectiveness of OWR by accounting for historical rainfall patterns and broadly aligning the policy to net lawn irrigation needs [8].

Purely technological approaches such as rain sensors and smart irrigation controllers that are designed to adjust lawn sprinkler system run times depending on local weather conditions might seem to be obvious tools to prevent overwatering during the wet season. However, at present, there has not been a lot research on the effectiveness of these technological solutions outside of controlled research stations [9, 10], and OWR remain one of the most widespread lawn water conservation strategies. Therefore, there is a continuing need for research that informs OWR policies.

Several studies have examined different aspects of the effectiveness of OWR. For example, some studies have shown that information campaigns and enforcement activities increase the water savings realized from OWR [5, 8, 11]. A number of studies have examined the long-term effectiveness of OWR and have shown that behavior hardening can negate some of the reductions in lawn water use where OWR are in effect for long periods [8, 12]. Other studies, including Kenny et al. [13], have shown that mandatory water restrictions yielded better water savings than voluntary water restrictions. Existing research has also shown that implementing mandatory water restrictions may actually increase water usage above pre-restriction levels, particularly in times of drought [7, 14, 15]. All of the aforementioned studies indicate OWR, when implemented solely as a notice informing residents of allowed watering days, can leave potential water savings unrealized and demonstrate the value of refining either the policy itself or the implementation of the policy to maximize water savings.

One way to refine OWR policy is to better match watering schedules with patterns of net irrigation need. While the standard of a day of the week lawn irrigation schedule facilitates compliance and enforcement, such schedules are indifferent to temporal changes in the weather which can greatly change the need to water the lawn. Consequently, because unvarying OWR schedules allot the same watering opportunity regardless of how much recent rainfall there has been, they can unintentionally promote overwatering when the lawn does not need it. OWR can also promote a ‘set and forget’ mentality among homeowners with automatic sprinkler systems to focus their attention to the legal schedule rather than directing that attention to the natural availability of water from rainfall, which can greatly offset or eliminate the need to irrigate. Thus, compliance does not necessarily equate to actual water conservation. This underscores the value of examining historical records of weekly weather data that can be used to inform OWR schedules [8, 16, 17].

One region where there is an opportunity to better align OWR policy with climate to improve lawn water conservation is South Florida. Similar to other metropolitan areas with a distinct wet/dry seasonal pattern, roughly 70% of the annual precipitation (~ 150 cm) in South Florida falls in the wet season from late May to mid-October [18,19,20]. Currently, Florida statute prescribes a permanent, year-round, regional water restriction policy in South Florida that allows weekly lawn irrigation on 2 or 3 days regardless of rainfall [21]. Such policy runs the risk of promoting overwatering behavior during weeks where rainfall alone is providing enough water for lawns. Thus, OWR can have the potential unintended consequence of sanctioning unnecessary water use. The example of South Florida demonstrates the importance of developing a connection between regional climate and lawn water use by highlighting and communicating the amount of water that rainfall typically contributes towards weekly lawn water demand at different times of the year [8, 16, 17].

In this paper, we present a climate-based approach to optimizing water savings from weekly lawn irrigation conservation strategies by analyzing regional historical patterns of rainfall and ET to determine the variability of net irrigation needs on a weekly basis throughout the year. We then use this weekly approach to historical rainfall and ET to compare the potential effectiveness of various water restriction schedules in a demonstration area. Our aim for this research was to develop a method for better estimating which weekly restriction schedules are best suited for a given locality using data that are readily available to local water managers. The demonstration area for this project was Palm Beach County, a large urbanized region in South Florida. The objectives were to (1) compile historical weekly rainfall and ET data to examine the temporal variability in rainfall, ET, and lawn water demand, (2) identify historical patterns in long-term average rainfall and ET to identify weeks of the year when rainfall alone typically meets lawn water demand, and (3) demonstrate the policy implications of various water restriction schedules by comparing the maximum allowed use under each policy to the net irrigation needs over a historical 20-year period.

Study Area

The study area consists of a developed 1672 km2 region of South Florida comprising the eastern urban gradient of Palm Beach County (Fig. 1). From north to south, it forms a narrow strip of development that stretches roughly 70 km long by 14 km wide, bordered by the Atlantic Ocean to the east and the Everglades wetlands to the west. Through a visual examination of satellite imagery, this urban gradient was identified as the contiguous land area on the eastern edge of Palm Beach County where more than 50% of the land use cover is developed and includes the type of irrigated land cover associated with municipal public landscapes and/or residential areas. The South Florida Water Management District (SFWMD) maintains a historical database of radar-based rainfall and satellite-based ET estimates at 2 km × 2 km resolution for the entire study area. Thus, appropriate data were available for the study area to allow us to derive historical weekly patterns of rainfall and ET.

Fig. 1
figure 1

The study area within the urbanized gradient of South Florida

Full size image

This area was selected for the study because it is representative of the typical population density of greater South Florida and also because it contains a large area of irrigated urban and suburban land use. According to data from the Palm Beach County Property Appraiser’s Office, the area is estimated to contain approximately 291,000 single-family residences built on parcels of various dimensions ranging from 0.10 acre to over 10 acres. While there is a diversity of home and parcel sizes across the study area, the lawns, whether situated in public or private spaces, virtually all consist of one type of turf species: St. Augustinegrass (Stenotaphrum secundatum), a regionally popular warm season turf grass [22].

In terms of water resource sustainability, the most significant feature in this urban-suburban gradient is the ubiquitous lawnscape that dominates the land cover and represents the largest category of demand for the residential water supply [8, 23, 24]. Automatic in-ground sprinklers are the norm for lawn irrigation and are found in almost all residences throughout the study area. Although the municipal water supply extends into most of the suburban residential areas, many households in this region have private groundwater wells or have installed pumps and lines to privately supply surface water for irrigating lawnscapes [8, 25, 26].

Outdoor Water Conservation Policy in South Florida

Under current regulation, South Florida is subject to the Year-Round Landscape Irrigation Rule (LIR), promulgated by the SFWMD under Chapter 40E-24 [21]. This policy has established mandatory OWR that prescribe year-round irrigation schedules limiting lawn and landscape irrigation to no more than 2 days a week with some counties having a variance that allows no more than 3 days a week (Fig. 2). The LIR was designed to prevent gross water use inefficiency by those who might otherwise irrigate in excess of 3 days a week. Further, the LIR pro-actively serves to extend the limited lawn watering behaviors that had been enforced during earlier periods of regional drought. The SFWMD views the LIR as a fundamental step towards conservation and improving outdoor water use efficiency that leverages a demonstrable conservation ethic with the public [27, 28].

Fig. 2
figure 2

Weekly irrigation frequency limits imposed by the Landscape Irrigation Rule in South Florida. Modified from https://www.sfwmd.gov/community-residents/landscape-irrigation. Accessed 18 July 2016

Full size image

The functional strategy of the LIR is communicated via weekly schedules. For example, odd addresses can irrigate on Monday, Wednesday and Saturday, making it easy for people to remember and follow their assigned days of the week where lawn irrigation is legally allowed. However, these weekly irrigation schedules are fixed, with no variation to account for seasonal differences in rainfall. Thus, neither the LIR nor the water restrictions are based on the regional climate and consequently permit lawn and landscape irrigation for large portions of the year where, given typical seasonal rainfall, little to no supplemental irrigation would be necessary.

Methods

Historical Rainfall Data

To evaluate long-term seasonal rainfall patterns in the study area, daily NEXRAD radar estimated rainfall data for the period 1996–2015 were obtained from the SFWMD. NEXRAD, a Doppler radar product of the Weather Surveillance Radar (WSR-88D), uses radar technology to estimate rainfall to the nearest 0.254 mm (0.01 in.). In South Florida, radio wavelength reflectivity data is gathered from five radars located in Miami, Melbourne, Tampa, Jacksonville and Key West, each of which can measure the reflection from falling raindrops out to a distance of 230 km. The SFWMD maintains a contract with a vendor to obtain data at a spatial resolution of 2 km × 2 km. The end-of-month radar estimates available through SFWMD are run through a QA/QC process that includes using algorithms to adjust radar estimates to rain gauge data [29, 30].

We chose to use radar estimates of rainfall because, while rain gauges provide a wealth of data in terms of length of record, the network gauges are located almost exclusively in optimized field settings on the periphery of the urban area which are generally a significant distance from the densely developed residential areas. Given that rain events in South Florida can be dominated by spatially erratic tropical and convective processes that are known to deliver widely different amounts of rain within a small geographical radius, rain gauge point data has particular spatial limitations [20].

Observations from citizen networks with personal weather stations and rain gauges located within the urban area such as Weather Underground (Wunderground.com) and Community Collaborative Rain, Hail and Snow Network (Cocorahs.org) can offer spatial diversity in precipitation data; however, the length and availability of historical data from these sources is limited. Additionally, the radar estimated rainfall from the SFWMD was a gridded, long term dataset, with consistent data formatting throughout and thus required little to no formatting or sorting prior to conducting our historical analysis.

Determining Weekly Rainfall

Using the high-resolution historical NEXRAD radar rainfall dataset, we computed the aerial average rainfall for each individual week of the 20-year historical period in the study area. This provided 1040 weeks of rainfall data that we considered single data points for our analyses. We also calculated long-term average rainfall for each week of the year (1–52) across the historical period. This analysis of long-term average weekly rainfall allowed us to discern seasonal weekly trends in rainfall throughout the typical annual cycle, while the data for all the 1040 individual weeks enabled us to examine the year to year variability in rainfall.

Historical Evapotranspiration Data

The SFWMD produces and maintains a historical database of satellite-based estimates of both reference evapotranspiration (RET) and potential evapotranspiration (PET) with the same spatial and temporal resolution as the NEXRAD rainfall estimates. Because RET refers to the evapotranspiration from a specific crop type rather than being generic, it can more closely approximate lawn water demand, making it better suited for a lawn irrigation water use study than potential evapotranspiration [31]. To derive estimates of RET, the SFWMD utilizes data obtained from the Geostationary Operational Environmental Satellites (GOES), then calibrates the data to correct for season and cloudiness biases using a modified form of the Priestly-Taylor algorithm and uses a short grass reference crop [32]. These satellite-based estimates of daily RET were obtained from the SFWMD for the study area for 1996–2015 to match the same time period of the rainfall data.

It is worth noting that urban settings are fundamentally different from the open field conditions where estimates of RET are based on measurements of solar radiation, temperatures, and wind speeds that are optimized by large, undeveloped, unshaded expanses of vegetation [33]. Just as the built and landscaped suburban environment provides wind breaks to reduce rainfall measurements, so does it also introduce wind breaks and shade factors that reduce rates of evapotranspiration [34, 35]. Consequently, we believe that the satellite-derived RET values obtained from the SFWMD may systemically overestimate the water loss that would be representative of the typical lawnscape, and our resulting calculations may overestimate lawn water demand.

Crop Coefficients

To further constrain the RET data and more closely estimate lawn water demand from study area lawns, we used monthly crop coefficients (Kc) to account for species-specific turfgrass water loss. By using monthly Kc values appropriate for South Florida, we could constrain our estimates of RET by estimating crop-specific evapotranspiration (ETc) from the urban lawn surfaces across the study area. There are a number of sources for Kc of different species of warm season turf grass [36]. We selected monthly Kc values developed by Romero and Dukes [37], modified from Stewart and Mills [38], that were based on the St. Augustinegrass turf species and tested in a South Florida location (Table 1).

Table 1 Kc values
Full size table

Determining Lawn Water Demand

We took the value of daily crop ET (ETc) to be roughly equivalent to the daily value for lawn water demand. Using the daily satellite RET data from the SFWMD and our chosen monthly Kc values, we calculated a daily ETc after Allen and others [39] as follows:

$$ {ET}_c={K_c}^{\ast }\ RET $$
(1)

where ET c is daily turf evapotranspiration (cm/day), K c is the monthly crop coefficient that corresponds to the given day of interest (Table 1), and RET is reference evapotranspiration (cm/day). Then, we estimated average weekly lawn water demand (ETc) for the entire study area using the same weekly approach as with the rainfall data. We determined both the aerial average of weekly ETc for each week across the 20 years of data and the long-term average ETc for each week of the year (1–52).

Determining Net Irrigation Needs

To determine historical weekly average net lawn irrigation needs in the study area, we compared our estimates of average weekly rainfall to average weekly ETc. We used the long-term averages for each week of the year (1–52) to identify the times of the year in which supplemental lawn irrigation has, on average, historically been unnecessary because the average weekly rainfall has typically either met or exceeded the average rate of ETc. We also determined how much supplemental irrigation has been, on average, historically necessary during times of the year when rainfall alone did not meet the lawn water demand. As with the other parameters, we also examined the data for all individual 1040 weeks of the study period to evaluate the year to year variability in net irrigation need.

Determining Policy Implications of Various OWR Schedules

We used our historical weather data set to broadly assess several different OWR schedules (Table 2) by comparing the maximum water use allowed by the policies (Umax) to the net irrigation needs during the historical study period. We determined the percent of time that the Umax of each policy would have been equal to or greater than the net irrigation need (resulting in a lawn water surplus) or less than the net irrigation need (resulting in a lawn water deficit). We also estimated the magnitude of those surpluses and deficits.

Table 2 Summary of the water restriction schedules evaluated in this study
Full size table

Since the idea of OWR is to define an upper limit on lawn water use, we assumed a maximum scenario where all households would irrigate with automatic sprinkler systems on their scheduled days. We drew upon estimates of per household lawn water use in a case study by Survis and Root [8] who measured the typical single-family household water use per lawn watering event during the late summer and fall in Wellington, Florida, a municipality within the study area (Fig. 1). This prior study [8] found that the typical single-family household in the study area had an estimated application depth of 1.89 cm per watering event with a standard error of 3.6%. Both the method and the estimates of lawn water use on a per cycle basis from automatic irrigation systems are supported by irrigation audits performed as an extension of local and regional water conservation initiatives through the US Department of Agriculture [40]. Thus, the Maximum allowed use (Umax) for each of the policies shown in Table 2 was calculated as follows:

$$ {\mathrm{U}}_{\mathrm{max}}=1.89\ {\mathrm{cm}}^{\ast }\ \mathrm{watering}\ \mathrm{days}\ \mathrm{per}\ \mathrm{week}\ \mathrm{allowed}\ \mathrm{by}\ {\mathrm{policy}}^{\ast}\#\mathrm{weeks} $$
(2)

We note that Umax is greater than actual lawn water use as some people may follow schedules where others may not. Additionally, our estimated application depth was based on measurements made during the hottest times of the year [8]. Even though there are a lot of assumptions and estimations in this approach, we demonstrate that a simple rough approximation such as this can be valuable for informing water restriction policy. Given that water managers are commonly working with scant data on water use, this simple approach that requires only an estimated application depth per watering event and a historical record of climate parameters, both of which are easily obtainable, offers a practical option for evaluating policy.

Results

Long-Term Average Rainfall Patterns on a Weekly Basis

The strong seasonal pattern of rainfall in urbanized South Florida comes into sharp focus when the historical rainfall data are analyzed with a weekly approach (Fig. 3). The typical year begins with some of the driest weeks, where average weekly rainfall was less than 0.76 cm. While low average weekly rainfall prevailed in the early dry season, there was a notable pattern of a cluster of heavy rain days that typically occurred in early March, between weeks 10 and 11.

Fig. 3
figure 3

Weekly approach to the natural water balance in South Florida. Historical average weekly rainfall is shown in relationship with average weekly lawn water demand (ETc) for the study area in South Florida. Note the duration of the wet season rainfall surplus relative to lawn water demand

Full size image

The wet season typically started dramatically in May between weeks 19 and 21, with average weekly rainfall jumping to above 3.81 cm. The wet season then tapered off between weeks 40 and 42. While wet season average weekly rainfall was 4.34 cm, the data also revealed a mid-summer dip in weekly rainfall around week 30 with an average of 3.02 cm of rainfall for that week. The wettest weeks of the year occurred around the first week of September in weeks 36 and 37 where late summer storms pushed weekly averages to more than 5.50 cm.

The transition from wet to dry season began in mid-October in week 43 where average weekly rainfall started to decline. There was generally an uptick in rainfall during week 45 where average weekly rainfall jumped to 3.20 cm before leveling off to around 1.65 cm for the remainder of the late dry season weeks.

Long-Term Average Lawn Water Demand (ETc) Patterns on a Weekly Basis

Lawn water demand (ETc) followed a somewhat smoother but still distinct seasonal pattern, roughly inverse to the pattern of average weekly rainfall (Fig. 3). The historical early dry season began in January and ended in mid-May, spanning weeks 1–20. Lawn water demand was typically lowest in December and January. Based on the historical average, at the beginning of the year ETc is around 1.24 cm and steadily increases through mid-May. In the typical year, average weekly lawn water demand progressively rises to a peak above 3.50 cm in mid-May (week 20). The historical data reveal that the weeks in the late spring, around week 20, had the highest lawn water demand.

By week 21, once the wet season was underway and average weekly precipitation soared above 2.54 cm, there was a steep decline in weekly lawn water demand. Average weekly ETc plummeted at the beginning of the wet summer season (week 21) following a downward trajectory throughout the summer and into the fall that continued to the end of the year as rainfall became lighter and more erratic. The lowest average weekly value for ETc occurred in late December (week 52) at 1.18 cm.

Long-Term Average Patterns of Net Irrigation Needs on a Weekly Basis

By comparing the patterns of rainfall to ETc throughout the average year on a weekly basis, the resulting seasonal pattern of net irrigation need demonstrates that average wet season weekly rainfall, by itself, provides more water than is necessary to sustain residential lawns and landscapes (Fig. 3). Average weekly rainfall alone provided 54% more water than was required by the typical lawn during the wet season. The late spring weeks were when, on average, the net irrigation need was the highest as the lighter rainfall progressively lost ground to the mounting ETc that peaked in May. The late dry season data displayed a different pattern. Here, average rainfall was adequate in many weeks but dipped below the average water demand of the lawn in portions of late October, November and late December.

Comparing Water Restriction Schedules

While long-term averages are useful for highlighting seasonal trends and identifying the typical times of the year when less supplemental lawn watering is needed, they can obscure year-to-year variability in the data. For example, wet season average weekly precipitation can be heavily influenced by storms that push the average rainfall well above what it is in a typical year. Conversely, periods of drought can skew the average weekly ETc to higher than typical values. In our historical climate dataset, the standard deviations in the precipitation and net irrigation need are greater than the long-term average precipitation and net irrigation need, indicating that there were several years during the study period when the long-term averages were not representative of typical conditions in the study area. Thus, after using the long-term average data to identify seasonal opportunities to save water, we evaluated the effectiveness of several different OWR schedules by comparing weekly Umax to weekly net irrigation need for all 1040 weeks of the historical study period.

This more detailed analysis revealed that the Umax for a 3-day week policy in addition to precipitation would have met or exceeded lawn water demand in the study area in all 1040 weeks of the historical study period (Fig. 4a and Table 3). If a homeowner had applied the Umax for a 3-day week policy, the lawn would have received an average of 6 cm/week more water than it needed. More than 75% of this surplus would have come from lawn irrigation and the rest from rain. We estimated that the cumulative surplus for the entire 20-year study period for a household with a 3-day Umax would have been 6300 cm. Thus, based on an estimated median US lawn size of about 550 m2 [41], the combination of Umax lawn watering and precipitation would have provided the typical lawn with over 35,000 m3 more water than it needed during the 20-year study period. About three quarters of this surplus would have been due to lawn irrigation. Thus, we estimate that over a 20-year period a 3-day policy in the study area could result in up to 27,000 m3 of unnecessary lawn water use. We recognize that this is probably an overestimate because lawn water use is probably less than Umax. However, our simple analysis using readily available data clearly shows that there could be significant additional water savings if steps were taken to better align the current 3-day policy with historical climate data.

Fig. 4
figure 4

Percent of years in the 20-year historical record that the Umax would have been greater than net irrigation need (surplus) and less than net irrigation need (deficit) for a 3-day OWR policy, b 2-day OWR policy, and c 1-day OWR policy

Full size image
Table 3 Quantitative summary of the surpluses and deficits that would have resulted from a household watering at the Umax of specific OWR schedules throughout the 20-year historical record
Full size table

Our analysis demonstrates that the Umax of a 2-day policy would have provided adequate water to the lawns during 99% of the 1040 historical weeks (Fig. 4b and Table 3). This result further emphasizes that the current 3-day policy does not adequately account for the significant contribution that rainfall makes towards lawn water demand in the study area. The only times that the Umax of a 2-day policy would not have fully satisfy net irrigation demand were during the late spring when ETc climbs and the heavy rains of the wet season have not begun (Fig. 3). However, even during this time of maximum net irrigation demand, we estimate that the largest weekly deficit resulting from the Umax of a 2-day policy during the entire 20-year record would have been 0.5 cm. Under a 2-day policy with Umax watering the cumulative deficit for the entire historical period would have only been 3.3 cm compared to a cumulative total surplus of 4300 cm. About 70%, 3000 cm, of this surplus would have come from lawn watering. Thus, for our median US lawn size of 550 m2, we estimate that watering at Umax under a 2-day policy could have resulted in up to 16,500 m3 of unnecessary lawn water irrigation over the 20-year study period.

Even the Umax of a 1-day policy would have met net irrigation need in the majority (81%) of the 1040 historical weeks and would have resulted in a significant cumulative surplus (1900 cm) over the historical period (Fig. 4c and Table 3). The portion of the surplus due to lawn watering as opposed to precipitation with a 1-day policy would have been 1140 cm, which would amount to 6270 m3 for a median US house over the 20-year study period. A significant amount of this overwatering would have occurred during the winter months when lawn water demand is at its lowest. This finding demonstrates that there is opportunity to conserve water, even during the dry season, by better aligning policy with long term historical net irrigation needs. However, because lawn water demand would not have been met in nearly 20% of the historical record and because weekly deficits would have been as great as 2.4 cm under a 1-day policy, such a policy would have probably resulted in some browning of lawns during parts of the 20-year study period.

That this 1 day a week policy would have allowed overwatering during the majority of the study period and yet would have resulted in browning lawns during some portions of the study period, highlights the major complication of developing OWR schedules: if OWR schedules allow for enough watering to meet lawn water demand during drier than normal periods, they will also allow for significant overwatering during periods with normal or above-normal precipitation.

Our analysis of historical climate data and evaluation of how various OWR policies align with regional climate trends demonstrate the benefits of considering historical climate patterns when designing OWR schedules. Our results clearly show that the 3-day policy does not adequately account for precipitation’s contribution to lawn water demand and that nearly all lawn water demand for the study area can be more than met by a 2-day policy. However, while even a 2-day and 1-day policy allow for overwatering, a 1-day policy would not have been adequate to meet net irrigation needs for 20% of the historical record. This demonstrates that OWR can only go so-far in helping to conserve water and also demonstrates the potential value of add-on policies that would promote turning off sprinkler systems and discourage lawn watering during wet periods.

Discussion

Identifying Seasonal Opportunities to Save Water

By design, weekly water restriction schedules ignore rainfall as the primary input of water to the lawn and thus miss the opportunity to frame lawn irrigation as a supplemental activity to make up for rainfall deficits. This is particularly true in areas where such water restriction schedules remain in place for long periods or are in effect year-round. The significance of this looms large for urban areas with seasonally wet climates such as South Florida. Here, while weekly irrigation schedules allow for adequate watering during the dry season, they may tacitly approve of overwatering during the wet season where rainfall greatly modifies the need to irrigate the lawn for weeks or even months at a time. Similarly, weekly schedules may allow overwatering even during winter months when lawn water demand is at its lowest. When water restrictions are unconnected to the natural water balance, they offer the public little information to maximize possible reductions in lawn irrigation practices and consequently leave considerable water savings on the table. Additionally, such longstanding and restrictive day of the week lawn water conservation strategies may lead people to believe that they are doing their part to help conserve water by irrigating only on sanctioned days of the week while they are actually overwatering. While the misalignment of policy to climate is obvious in the wet season in South Florida, the findings of this study show that even a 1-day a week schedule can contribute to an overuse of water resources during the winter weeks in December, January, and early February when lawn water demand is at its lowest (Fig. 4). These are the times of year where a 3-day or 2-day OWR is least effective as a water conservation strategy.

There is a clear need to find ways to develop OWR schedules that are simple to implement and comply with but that are also aligned with the natural variability in lawn water demand which govern net irrigation needs. This study demonstrates how a relatively simple and low-cost analysis of historical trends in average weekly rainfall and ETc can be used to assess the natural backdrop of water availability from rainfall. When combined with a parameter, such as Umax, that is easy to calculate with readily available data, water managers can evaluate the “fit” between the most common weekly water restriction policies and local or regional weekly net irrigation needs to help managers weigh the benefits and drawbacks of different OWR schedules (Table 4).

Table 4 Effectiveness of OWR schedules during historical study period
Full size table

The Environmental Cost of Unnecessary Lawn Irrigation

Our projections of Umax were based on research on single-family residences and did not consider the scale of water use that occurs in higher density housing, retail shopping areas, and municipal parks and medians. However, our analysis draws attention to the scale of overwatering that may be sanctioned by policies not synced to the long-term temporal variability of climate, and the volume of water that might otherwise be conserved in the environment as part of the natural water cycle. Unnecessary or over-irrigation, whether the source of water is municipal or private, comes at the expense of the natural system as excess cycling of water through urban turfgrass can also be viewed as a water quality issue. Lawn irrigation, when not necessary, displaces a large volume of water that partially becomes a contribution to urban runoff and abets the infiltration of surface pollutants such as fertilizers and pesticides to vulnerable surficial aquifers [23, 42].

Adopting water restriction guidelines that are more closely synced with the natural water balance would not only improve the effectiveness of lawn watering but would also significantly reduce withdrawals of water from natural systems. While the displaced water used for lawn irrigation is not entirely consumptive such as indoor water uses are, a substantial portion of the water applied by sprinkler systems is depleted by evaporation in addition to losses from transpiration, interception, and runoff and is therefore not available to contribute to the recharge of local aquifers [20, 26]. Given the scale of water withdrawals attributable to residential lawn irrigation, this represents a significant and often unnecessary disruption of the urban water cycle.

Conclusions

The weekly analysis of rainfall and ET data and the example we have provided with its application to assessing OWR policy in South Florida represents a simple and relatively inexpensive tool to help managers develop policies that explicitly incorporate climate-water interactions. In South Florida, the permanent year-round water restrictions represent a fundamental step in the direction of responsible use of water resources. Now that the idea of limiting lawn watering to 2 or 3 days a week on a non-drought basis has been well established, it may be time to transition policy to address the next step towards true water conservation based on weather based inputs. Thus, evaluating trends of historical rainfall and lawn water demand on a weekly and seasonal basis is necessary to optimize the design of outdoor water restrictions and better couple human activities with the natural system. While our study took place in South Florida where permanent year-round OWR are in place, areas where non-permanent OWR are implemented for shorter durations or in times of drought could employ an analysis of net irrigation needs to develop optimized watering schedules. Other studies suggest there are opportunities to save water by better calibrating landscape irrigation with climate to reduce overwatering even in arid regions [43,44,45,46].

Additionally, given that all three OWR policy schedules we evaluated sanction some degree of overwatering, this suggests that relying solely on limiting the frequency of lawn irrigation might not be enough to realize maximum possible water savings. There remain significant opportunities to improve water conservation via multi-pronged approaches such as real-time weather-based information campaigns that discourage watering when rainfall alone has met or exceeded lawn water demand. This type of strategy can enhance the effectiveness of existing OWR with climate-optimization [5].