Abstract Water tables are dropping by approximately one meter annually throughout the North China Plain mainly due to water withdrawals for irrigating winter wheat year after year. In order to examine whether the drawdown can be reduced we calculate the net water use for an 11 year field experiment from 2003 to 2013 where six irrigated crops (winter wheat, summer maize, cotton, peanuts, sweet potato, ryegrass) were grown in different crop rotations in the North China Plain. As part of this experiment moisture contents were measured each at 20 cm intervals in the top 1.8 m. Recharge and net water use were calculated based on these moisture measurement. Results showed that winter wheat and ryegrass had the least recharge with an average of 27 mm/year and 39 mm/year, respectively; cotton had the most recharge with an average of 211 mm/year) followed by peanuts with 118 mm/year, sweet potato with 76 mm/year, and summer maize with 44 mm/year. Recharge depended on the amount of irrigation water pumped from the aquifer and was therefore a poor indicator of future groundwater decline. Instead net water use (recharge minus irrigation) was found to be a good indicator for the decline of the water table. The smallest amount of net (ground water) used was cotton with an average of 14 mm/year, followed by peanut with 32 mm/year, summer maize with 71 mm/year, sweet potato with 74 mm/year. Winter wheat and ryegrass had the greatest net water use with the average of 198 mm/year and 111 mm/year, respectively. Our calculations showed that any single crop would use less water than the prevalent winter wheat summer maize rotation. This growing one crop instead of two will reduce the decline of groundwater and in some rain rich years increase the ground water level, but will result in less income for the farmers. Figures Fig 4Fig 5Fig 6Table 1Fig 1Fig 2Table 2Fig 3Table 3Fig 4Fig 5Fig 6Table 1Fig 1Fig 2 Citation: Yang X, Chen Y, Pacenka S, Gao W, Zhang M, Sui P, et al. (2015) Recharge and Groundwater Use in the North China Plain for Six Irrigated Crops for an Eleven Year Period. PLoS ONE 10(1): e0115269. doi:10.1371/journal.pone.0115269 Academic Editor: Juergen P. Kropp, Potsdam Institute for Climate Impact Research, GERMANY Received: September 5, 2014; Accepted: November 6, 2014; Published: January 27, 2015 Copyright: © 2015 Yang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited Data Availability: All data supporting the results of this article is fully available without restriction in Cornell University’s permanent eCommons repository (http://hdl.handle.net/1813/37320). Funding: The authors are grateful for grants from the National Key Technology R&D Program of the People’s Republic of China (2011BAD16B15 and 2012BAD14B03), the China Scholarship Council for Xiaolin Yang (Grant No. 201306350107) and the Chinese Universities Scientific Fund (2013YJ001) and the support of College of Agriculture and Life Sciences at Cornell University by providing the use of their facilities free of charge. Competing interests: The authors have declared that no competing interests exist. Introduction Available water resource has become scarcer especially in arid and semi-arid regions, leading to mining of groundwater in many places [1]. Since agriculture is the largest water user and the increasing and more affluent population consumes more milk and meat products, one of the greatest challenges for the agricultural sector is to produce more food with less water [2]. This is especially critical in the North China Plain (NCP) where agricultural sustainability is threatened by the continued decline in the region’s groundwater table. The North China Plain is the premier agricultural area in China where 61% of the nation’s wheat, 45% of the maize, 35% of the cotton and 64% of the peanuts are produced [3]. In this area, crop productivity relies heavily on irrigation from groundwater because nearly all usable surface waters are stored in reservoirs for transfer to metropolitan areas [4–6]. Groundwater, and to a limited extent the city waste water downstream of large cities, are used for three to four flood irrigations per year with 300 mm or more water in total [7–10]. Using a large scale weighing lysimeter, Liu et al. [11] reported annual agricultural water consumption for the NCP of about 800 to 900 mm. This is much more than the regional precipitation which averages 480 mm/year (yr) [12]. As a result, crops are irrigated using groundwater as the only available source. The groundwater tables have fallen at a rate of 1m per year in the last 20 years which has led to severe land subsidence in some areas [13, 14] and salt intrusion in areas near the coast. Earlier regional studies of irrigation efficiency in the North China Plain have used several methods to calculate recharge and evapotranspiration [8, 15–27]. Estimates of actual evapotranspiration (ETa) of winter wheat varied widely with a low of 346 mm/yr [19] to a high of 737 mm/yr [25] and the ETa of summer maize from a low of 280 mm/yr [27] to a high 374 mm/yr [22]. Ma et al. [18], Liu et al. [21], Zhang et al. [22], Kong et al. [24] and Zhao et al. [28] used a soil water balance method to calculate ETa using precipitation and irrigation as input and neglecting deep percolation to below the soil profile. Thus in this method any deep percolation is counted within ETa. Other researchers accounted for recharge in their simulation models. Sun et al. [8] simulated recharge as a constant fraction of both irrigation and effective rainfall (e.g. Luangcheng County Water Policy and Integrated Water resources Management Office, 1993 [29]) for the period of 1999 to 2002 and found for the NCP a recharge of 25–37 mm under winter wheat in a period including 309–405 mm total irrigation and 347–402 mm precipitation. Ma et al. [27] quantified drainage out of the root zone with a simplification of Darcy’s law, calculating a recharge of winter wheat at 5mm and 64 mm in 2007 (300 mm irrigation and 83 rainfall) and 2008 (210 mm irrigation and 186 rainfall), respectively. Liu et al. [26] calculated percolation from the root zone based on the relation between unsaturated water conductivity and soil moisture content at the bottom of the soil profile, and reported there was 1–18 mm percolation during the winter wheat growing season from 2008 to 2012 with irrigations totaling 226–304 mm, using a root zone was of 1.5 m. However, they did not report the recharge of other crops. Kendy et al. [15, 16] validated a one—dimensional soil—water—balance model assuming that drainage from the soil profile was continuous to estimate areal recharge from winter wheat-summer maize and reported a recharge between 71–170 mm for the whole rotation from 1998 to 2001 under irrigation of 210–328 mm and annual precipitation of 347–402 mm. Using Kendy’s model, Zhang et al. [17] simulated recharge amounts varying between 33–87 mm for winter wheat from 1998 to 2001 with 362–405 mm total irrigation and 53–134 mm precipitation in its growing season. Meanwhile, he also reported that recharge for summer maize was 127 and 12 mm in 2000 (126 mm irrigation and 348 mm precipitation) and 2001 (67 mm irrigation and 212 mm precipitation), respectively. Finally Hu et al. [12] indicated that, based on a combination of SWAT and MODFLOW models, a 39% reduction in irrigation pumping for winter wheat-summer maize would induce groundwater recovery and restoration to the pre-development hydrologic conditions of 1956 in about 74 years in the NCP. As a simplifying assumption, all of the above studies on irrigation efficiency assume either implicitly or explicitly that there is no surface runoff. To neglect surface runoff on the large scale in this area is justified. Since 1980, after the reservoirs were built to collect the river flow from the mountains for industrial and domestic water supply, rivers were dry except in the high rainfall years of 1988 and 1996 [30]. On a small scale, runoff is unlikely due to the high infiltration capacity of the soil and flat topography allowing water to stand in small puddles until it infiltrates or evaporates [12]. In the North China Plain there is no surface water available for irrigation and groundwater is the only source for irrigation. It is useful to know the net ground water use, namely the difference in irrigation water withdrawal and recharge. The excess of irrigation withdrawal above recharge year after year drives the cumulative groundwater level decline. Kendy et al. [15] found that on average, when growing both winter wheat and summer maize, net water use was about 125–212 mm per year from 1998 to 2001. (In this summary, positive values represent greater irrigation than recharge, which is a negative effect on the groundwater level.) Zhang et al. [17] showed net water uses of winter wheat-summer maize from 181–321 mm between 1998 and 2001 in the NCP. Sun et al. [8] reported net water use for winter wheat ranging from 230 mm to 251mm from 1999 to 2002 in the NCP. Sun et al. [31] calculated that the net groundwater use required for winter what-summer maize with reduced input of water and nitrogen under farmers’ practice still surpassed 300 mm yr-1 in the NCP. Liu et al. [26] reported that the net water use in the growing season of winter wheat ranged from 226 mm to 292 mm from 2008 to 2012 in the NCP. Based on a root zone model, Ma et al. [27] reported the net water use for winter wheat by itself in the NCP was 146 mm in 2007 and 295 mm in 2008, respectively. However, the models were not directly validated leaving a large uncertainty about recharge and net water use amounts. The exception was the model of Kendy et al. [15, 16] that was validated against observed soil moisture data for the winter wheat-summer maize rotation. They used lysimeter data published by Liu et al. [11] from the Luancheng experimental station site. While these cited reports are very helpful, crop rotation and longer term field research on the net water use and recharge of the staple crops in the NCP remains scare. Because of limited knowledge about recharge and net water use mentioned above in the NCP, an 11 year experiment was carried out that starting in October 2002 to study the moisture depletion patterns of six crops including winter wheat, summer maize, peanuts, cotton, sweet potato and ryegrass in five cropping systems. In this paper we analyze this experiment during which soil moisture content was measured in increments of 20 cm every 10 days to calculate the net water use of these crops as a function of crop rotation, irrigation schedule and precipitation. The overall objective of this study is to calculate the recharge and the net water use directly based on measured soil moisture content, pan evaporation, irrigation and precipitation for important crops in the North China Plain. Specifically, we will (i) quantify the deep percolation of the six crops from 2003 to 2013; (ii) re-express the difference between deep percolation and irrigation to infer groundwater table change; and (iii) calculate the net water use for the six crops from 2003 to 2013. Materials and Methods Ethics Statement This field experiment was conducted in the Luancheng Agro-ecosystem station (37°50’N, 114°40’E, altitude 50 m) which is a long-term experimental site and belongs to the Chinese Academy of Sciences. This research was performed in cooperation with China Agricultural University. The farm operations of this experiment were similar to rural farmers’ operations and did not involve endangered or protected species. This experiment was approved by China Agricultural University and the Chinese Academy of Sciences. Study site, soil and climate The Luancheng Agro-ecosystem station is one of 36 agricultural ecosystem stations in the Chinese Ecological Research Network (CERN). It is located in Luancheng County in Hebei Province. Its property and operation are representative of the agricultural production and climate conditions in the northern part of the North China Plain, where winter wheat-summer maize rotation is the main cropping system (Fig. 1). The water table in Luancheng County has been constantly declining since the 1970s, which fell from 11 m below the surface to about the current 42 m with an annual decline of approximately one m yr-1. The experimental site has a warm temperate zone, semi-humid, monsoon climate. The annual mean air temperature is 12°C. The average annual rainfall over the last 20 years was 480 mm, with sharp yearly fluctuations and an erratic seasonal distribution. Generally, 60–70% of the yearly precipitation occurs from June to August. The monthly precipitation and averaged temperature are shown in Fig. 2. The frost-free period is about 200 days from April to October. The experimental site has a sandy loam in the surface layers, light/medium loam at a depth of 40–80 cm, and light clay below 80 cm. The soil profile properties are given in Table 1[32]. The plow layer was about 0.2 m thickness, contained 11 g/kg of organic matter, 1g/kg of total nitrogen, 36 mg/kg of available phosphate, and 96 mg/kg of available potassium. The soil pH is 7.8 [33]. thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Table 1. Soil properties at Luancheng experiment site before the start of experiment. http://dx.doi.org/10.1371/journal.pone.0115269.t001 thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 1. Location of Luancheng County and the North China Plain. http://dx.doi.org/10.1371/journal.pone.0115269.g001 thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 2. Temporal distribution of monthly precipitation and temperature from 2003 to 2013 at Luancheng experiment site. http://dx.doi.org/10.1371/journal.pone.0115269.g002 Experimental design and crop management In October 2002, five cropping systems were established on 15, 4 × 7.5 m plots in a randomized complete-block design with three replicates. Between two plots, there was a 1 m-wide zone without irrigation to minimize cross-plot effect. The five cropping systems were (1) winter wheat—summer maize (WS; 1-year cycle), (2) peanuts→winter wheat—summer maize (PWS; 2-year cycle), (3) rye—cotton→ peanuts→ winter wheat—summer maize (RCPWS; 3-year cycle), (4) sweet potato→ cotton→ sweet potato→ winter wheat—summer maize (SpCSpWS; 4-year cycle), and (5) continuous cotton cropping (Cont C). Planting, harvesting, fertilizers and irrigation applied were the same for each crop independent of its rotation and according to local agronomic practices (Table 2). Fertilizer N (primarily urea fertilizer) was broadcast prior to seedbed preparation and top applications of N, P and K fertilizer were made during the crop growth period. Data were analyzed up to October 2013. thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Table 2. Experimental treatment of different crops. http://dx.doi.org/10.1371/journal.pone.0115269.t002 Data collection Average daily air temperature, precipitation and pan evaporation were obtained from an automatic weather station located 100m from the experimental plots. The groundwater table was measured about 400 m from the experimental plots at the Experimental Station’s well. Soil volumetric moisture content was measured by a neutron probe (L520) at increments of 20 cm every 10 days to a depth of 180 cm. The top 20 cm soil moisture was measured by a thermo-gravimetric method because neutron probe measurement is inaccurate near the surface. Calculations and measurements Aquifers in the North China Plain are potentially recharged by deep percolation from excess precipitation and irrigation [12, 15, 34]. In order to calculate the cumulative percolation, we calculated the water balance of the top 180 cm for 10 day intervals between consecutive measurements of soil moisture content. Any water shortage in the top 180 cm after subtracting the estimated evapotranspiration by the crop is considered deep percolation which eventually becomes recharge to the groundwater. This assumes implicitly as in other studies within the North China Plain that surface runoff can be neglected [12, 34], which was confirmed by direct observation. It also assumes that interflow is small because the horizontal hydraulic gradient is close to zero in the nearly flat alluvial plain. Finally, with the local groundwater table much deeper than 4 m below the ground surface, capillary rise is negligible [8]. Thus, by neglecting lateral flow and water moving upward from the groundwater, the water balance equation for the deep percolation between moisture measurements for each crop can be written as: (1) Where the storage S is as measured in the top 180 cm of the soil profile; S(t-Δt) is the storage in the profile at the previous measurement time; Δt is the time between measurements which was usually 10 days; DP is the average daily percolation for the interval (mm/day); P is the daily precipitation (mm/day); I is the daily irrigation (mm/day); and ETa is the estimated daily actual evapotranspiration (mm/day). P, I and ETa were summed over Δt days. ETa was calculated with the FAO method [35], e.g. (2) Where Kc is the crop growth coefficient; Pc is a pan coefficient that converts the evaporation from a pan to potential evapotranspiration from a well-watered, fully leafed crop; and ETp is the measured pan evaporation. The crop coefficients Kc varys during the different growth stages of the crop [35]. Here, Pc was equal to 0.7 which was the same as Kendy et al. [15] at the same site. Kc was obtained from the study of Liu et al. [11] that determined values for the Luancheng Station site using lysimeters (S1 Table). In cases when the deep percolation in Eq. (1), DP, was computed to be negative, we assumed that the FAO method (which gives potential evapotranspiration from a very moist surface) overestimated the actual evapotranspiration. Then we reduced the ETa so that DP would become zero in Eq. (1). Net water use, NWU, by the crop from the perspective of the groundwater can be estimated based on DP and pumpage for irrigation. The meaning of a negative “use” is that storage in the aquifer has been reduced; a positive “use” means that aquifer storage increased (3) Where n was the number of days from sowing till harvest. Statistical evaluation Analysis of variance (ANOVA) was carried out using Statistical Analysis System 9.3 [36] software, and differences were considered significant at the 0.05 level. Results and Discussion Deep percolation and evaporation The temporal distributions of precipitation and irrigation and the calculated percolation and evaporation during the period of 2003 to 2013 are presented in Fig. 3 for various crops in the SpCSpWS (sweet potato→ cotton→ sweet potato→ winter wheat—summer maize) rotation. Analogous information for the other four cropping systems is provided in S1 Fig. It is obvious that intensive rains in summer generate more deep percolation than the smaller precipitation events and irrigation applications during the remainder of the year. These findings are consistent with Kendy et al. [15] at the same experimental station and Bradbury et al. [38] in Wisconsin U.S. thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 3. Distribution of daily percolation and ETa with different precipitation and irrigation in SpCSpWS crop rotation during the period of 2003–2013. (Interval is the days between two measurements; SpCSpWS rotation: sweet potato→ cotton→ sweet potato→ winter wheat—summer maize). http://dx.doi.org/10.1371/journal.pone.0115269.g003 By summing the water losses between the soil measurements, the percolation and ETa for each crop can be calculated with Equations (1) and (2) (Table 3). Results revealed that cotton had significantly more percolation (51–359 mm) than all other crops (P< 0.05), particularly when it was grown as part of the cotton monoculture (Cont C). The reason for the large percolation amounts was the extra irrigation water given to cotton compared to other crops. Liu et al. [39] reported the extra irrigation was needed to meet the water requirement of cotton if there was not much rain. The high standard irrigations, together with the high rainfall in 2004 that fell in a relatively short time period at the beginning of the planting season on relatively wet soil resulted in the greatest percolation of 359 mm under the cotton crop. The minimum percolation for cotton was 51 mm in the following year (2005), mainly because of poor rainfall in that year. Low soil moisture carried over from the preceding cotton in late 2004 also contributed. Irrigation, as mentioned before, was based on a schedule that did not consider soil moisture content. thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Table 3. Deep percolation of different crops in five crop rotations during the growing season from 2003 to 2013 in the North China Plain (mm). http://dx.doi.org/10.1371/journal.pone.0115269.t003 In general, percolation amounts under winter wheat and ryegrass were the least of all crops (in some cases there was no percolation), due to low rainfall (rain stopped generally in the beginning of October) and low initial soil moisture content in the soil profile (Table 3). The long October-May non-precipitation winter wheat cultivation season relied on groundwater irrigation, which was the main reason behind saturated zone storage loss in this area. Some of the irrigation water that we applied would have built up in the soil in winter and spilled downward to the aquifer, but instead was transpired by the winter crops. Similarly to cotton, sweet potato had a relatively high percolation rate ranging from 25 mm to 136 mm. However, there was no percolation in the growing season of sweet potato in 2007, due to the poor precipitation and high evapotranspiration. Peanuts’ growing season percolation ranged from 7 mm to 335 mm (Table 3). Summer maize had the annual average percolation of 44 mm/yr. However, Percolation for all crops was small in 2007 and 2010 because rainfall was scarce (Table 3). Our results in Table 3 indicate that percolation of water under regionally irrigated cropland is common. This is contrary to the assumptions made by some of the early investigators calculating the crop water use [18, 22, 24, 25]. Therefore, in the next section we demonstrate that our predictions of percolation that recharges the groundwater are realistic and can represent the observed decline in the ground water tables in the North China Plain. Validation of deep percolation calculation In order to check the percolation calculation we re-express the difference between the calculated deep percolation and irrigation into infer the groundwater table change. Thus we compare our time series of calculated net water use with the time series of observed groundwater table change at the Luancheng experimental station. Fig. 4 shows the monthly observed water table at the experimental station along with our simulated values which will be explained shortly. thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 4. Comparison of the historical observed and simulated groundwater table for winter wheat-summer maize rotation at Luancheng site in the North China Plain. http://dx.doi.org/10.1371/journal.pone.0115269.g004 The depth of the groundwater table d can be predicted as (4) Where d is depth of the ground water in reference to the land surface at zero; R and I are the recharge and irrigation water withdrawn respectively in the period between the current time step, t, or the previous time step t-Δt, f is the fraction of the land under irrigation and η is the specific yield or drainable porosity. R, I and d are in meters. To calculate the recharge, since the groundwater table below the surface is at 30 m during the experiment period there is a notable time lag between when the water percolates out of the root zone and when it arrives at the groundwater. To estimate this time lag, we look at the ground water table historical behavior. We concentrate on a wet period starting with 1995 with annual rainfall of 510 mm/year; 1996 with 774 mm/year and then 1997 was dry with 272 mm/year. The ground water table as shown in Fig. 4 peaked in 1997 and 1998 indicating that there is one year time delay and the effect is spread over two years. Since our model is not refined enough to predict the inter-annual variation of the ground water we use a yearly time step. Based on this the average annual recharge is predicted as (5) Where the subscript y is the year and is the annual averaged areal deep percolation. In order to calculate the annual averaged deep percolation in the area around the Luancheng experiment station we note that 90% of the area around the Luancheng experiment station is irrigated and 85% is managed under the winter wheat-summer maize rotation [31]. Thus by summing the deep percolation under winter wheat and summer maize (Table 3), we can obtain the annual percolation near the experiment station. In addition based on these observations we assign a value of 0.85 to f in Eq. (4). Since the amount of irrigation is known from table 2 (based on farmer’s practices), the last parameter that need to be estimated is the specific yield in order to find the decline in ground water levels. Specific yield at the site has not been measured. Kendy et al. [15] assumed 0.2 and found a good fit between their computed deep percolation—irrigation and monitor well observations without considering the other factors above. Groundwater modeling literature covering the vicinity assumes 0.10–0.23 [40] and 0.12–0.18 values [12]. Standard texts often cite 0.2 for alluvial gravel and sand like that found beneath the Experiment Station. So we consider an 0.2 value to be reasonable. The predicted groundwater tables in our winter wheat-summer maize rotation are shown in Fig. 4 after the above calculation. It starts from the year of 2004. Due to this experiment initiated from October 2002, we set a percolation out of the root zone in 2002 the same as in 2003 by lack of a better estimate. As it can be seen in Fig. 4, the predicted annual average groundwater table follow the observed levels well, within the range of monthly variability of the observed water tables. Statistical measures indicate this as well with the root mean square error (RMSE) of 0.6418 m, the mean relative error (RE) of 1.35%, and Nash-Sutcliffe of 0.9633. These three criteria formula to quantify the deviation of the modelling results from the observed data were expressed in Ma et al. [27]. Shu et al., [41] reported among others that the excessive irrigation is generally assumed to be the key factor contributing to declining groundwater tables in the North China Plain. In contrary, especially Eq. (4) distinctly shows that the recharge nor the irrigation by itself are causing the decline of the ground water. It is the difference between the two that is responsible for rate that the ground water changes. Similarly recommendations to decrease the groundwater decline by either decreasing irrigation [12] or increasing the percolation [42] likely should be reconsidered. Net Water Use Exploring groundwater balance Eq. (4), there are two ways to reduce the aquifer depletion by reducing the net water withdrawal (and decreasing the net water use) or by decreasing the fraction of agricultural land (and minimizing the non-evaporation from cropping land). We next explore the choice of crop toward by minimizing the negative balance between irrigation, I, and recharge, R, using our experiment results. Thus to consider what crop is best for decreasing the reduction of groundwater, we calculated the net water use for each crop from the perspective of groundwater using Eq. (3) (Fig. 5). Negative values indicate that irrigation pumpage exceeded recharge, positive that recharge exceeded pumpage. From Fig. 5, the net water withdrawal of winter wheat was greatest among the tested crops with an average of 198 mm/yr in various crop rotations mainly due to the low precipitation in winter and continuous evapotranspiration for regular growth, drawing mostly upon 225 mm irrigation provided in winter wheat growing season. The ryegrass, winter cover crop, with irrigation of 150 mm in each growing season, also had high net water use with an average of 112 mm/yr due to ongoing evapotranspiration and low rainfall in the winter dry season. Under the same 150 mm irrigation, but in a wetter part of the annual weather cycles, sweet potato and peanuts had annual average net water use of 74 mm/yr and 32 mm/yr, respectively. Sweet potato had its highest net water use of 150 mm in 2007 due to the low growing season precipitation and no percolation. Peanuts in the PWS rotation had an extra recharge of 89 mm to the groundwater in 2013. The annual average net water withdrawal of summer maize from the groundwater reached 71mm/yr with the 105 mm irrigation, while in some years summer maize actually replenished the groundwater in the PWS rotation by returning 125 mm in 2004 and 48 mm in 2012. Cotton had the least annual average net water use of 14 mm/yr under its irrigation regime of 225 mm in each growing season. In wet years, cotton was the most likely crop to reverse the water extraction from the groundwater replenishing the aquifer by 18–134mm/yr (S2 Table). Overall, the annual average net water use decreased in the order of winter wheat > ryegrass > sweet potato > summer maize > peanuts > cotton (Fig. 5). thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 5. Distribution of Net Water Use (recharge minus irrigation) of six irrigated crops in the growing season from 2003 to 2013 at Luancheng experiment station in the North China Plain. (The box plots show the 25, 50, 75 percentiles. The squares and lines in the box plots indicate the mean and medium, respectively. The crosses indicate the minim und maximum. Negative data indicate the net groundwater depletion.) http://dx.doi.org/10.1371/journal.pone.0115269.g005 Across all crops, the sharp yearly fluctuations and the erratic seasonal distribution of annual precipitation contributed to a wide variation in any crop’s net water use. Our results are in agreement with Yang et al. [43] who estimated that the crop water requirement for five major crops (wheat, maize, cotton, fruit trees, vegetables) in NCP using crop models DSSAT and COTTON2K, and found winter wheat accounted for over 40% of total irrigation water requirement in the plain, while summer maize and cotton together accounted for 24% of the total irrigation water requirement. In order to find the relationship between net water use and rainfall and recharge, Fig. 6A plots the deep percolation (that eventually becomes recharge) versus the sum of precipitation and irrigation (P+I) over the growing season (from sowing to harvest) for each of the crops and for each of the years from 2003 to 2013. Although there is a general relationship in which the percolation increases with increasing P+I, there is a large variation between the various crops. The fit becomes much tighter in Fig. 6B when we take the duration of the growing period into account by dividing the total amounts in the growing season by the number of days of the growing period. Thus in Fig. 6B, we express the percolation versus P+I amounts by expressing the quantities in mm/day. Results reveal that there is a threshold, T, of around 3 mm/day. Above the threshold there is almost a linear relationship between P+I and percolation with a slope close to 1 (Fig. 6B). Below the threshold the percolation varies with the average percolation generally less than 0.75 mm/day with an average of approximately 0.2 mm/day. thumbnail Download: PPT PowerPoint slide PNG larger image TIFF original image Fig 6. Percolation as a function of precipitation plus irrigation (P+I) during the growing season (A) and daily percolation as a function of the daily (P+I) (B) for different crops in the North China Plain. (In the legend, the data of Sweet potato, Cotton, Winter wheat, Summer maize, Ryegrass and Peanuts come from this study. Winter wheat1 come from Sun et al., (2003) [8]; Winter wheat2 and Summer maize1come from Zhang et al., (2004) [17]; Winter wheat3 come from Liu et al., (2013) [26]; Winter wheat4 come from Ma et al., (2013) [27].) http://dx.doi.org/10.1371/journal.pone.011

The North China Plain is one of the most densely populated regions in the world. However, water scarcity is a growing problem. Action has been taken by The North China Plain Water Conservation Project, aiming to enhance beneficial use of water resources, agriculture production capacity, and farmer incomes. The key lesson is that approaches to water savings in agriculture are more likely to succeed if appropriate incentives are given to farmers. Description The North China Plain is one of the most densely populated regions in the world, encompassing the capital Beijing. The soil here is fertile, but water scarcity is a growing problem. In many areas, increasing agricultural demands can no longer be met by adding hydraulic infrastructure, and groundwater is over exploited. Water supplies to agriculture are under stress from competing demands for other uses. The problem of water scarcity is compounded by worsening water pollution from heavy industrialization and rapid urban population growth. Action taken The North China Plain Water Conservation Project financed by the World Bank (74 mill. USD) aimed to enhance beneficial use of water resources, agriculture production capacity, and farmer incomes. The project had four components: irrigation and drainage works and on-farm systems, including canal lining, low pressure pipes, drains, wells, surface irrigation improvements, sprinklers, and micro irrigation systems, agriculture support and services, including land leveling, non-tillage in the dry season, deep plowing in the rainy season, soil fertility improvements, organic and plastic mulching, cropping pattern adjustments, seed improvements, balanced fertilization and improvements to planting and cultivation techniques, forestry and environmental monitoring of the project’s impact on soil and water, and; institutional development and capacity building for water and soil conservation. Outcomes The sets of integrated water saving measures, including improvements in irrigation infrastructure, contributed to water savings while increasing agricultural yields and incomes. Among the project’s contributions are: Increased water productivity and reduced consumptive use. The value of agricultural production per unit of water consumed increased in the range of 60-80%, agricultural production tripled and farmer per capita incomes increased between 10-500%. More sustainable groundwater use. Groundwater depletion was reduced to negligible levels or eliminated. Strengthened institutions arrangements for irrigation system operation and maintenance. The original project target was 100 water user associations (WUA), but more than 500 were established. For the first time on this scale in China, WUAs assumed responsibility for both financing and operating irrigation systems. Water charges. Volumetric water charges were initiated. Importance of the case for IWRM Development of WUAs. The success of the WUAs stemmed for two principles: a) democratic self-organized associations based on hydraulic boundaries, and b) water measuring with corresponding water charges on a volumetric basis. Important factor was also inclusion of farmers form the beginning in the project, and the active support of both the Ministry of Water Resources and local governments. Importance of economic incentives. Approaches to water savings in agriculture are more likely to succeed if appropriate incentives are given to farmers to modify their practices. Monitoring and evaluation for technical innovations. An appropriate monitoring and evaluation system is necessary to verify efficiency of integrated water saving measures in agriculture, and to share the information with WUAs.

The North China Plain (NCP) is one of the most important agricultural regions in Asia and produces up to 50% of the cereal consumed in China each year1, 2. To meet increasing food demands without expanding croplands, annual agricultural practice in much of the NCP has changed from single to double cropping3, 4. The impact of double cropping on the regional climate, through biophysical feedbacks caused by changes in land surface conditions, remains largely unknown. Here we show that observed surface air temperatures during the inter-cropping season (June and July) are 0.40 °C higher over double cropping regions (DCRs) than over single cropping regions (SCRs), with increases in the daily maximum temperature as large as 1.02 °C. Using regional climate modelling, we attribute the higher temperatures in DCRs to reduced evapotranspiration during the inter-cropping period. The higher surface temperatures in June and July affect low-level circulation and, in turn, rainfall associated with the East Asian monsoon over the NCP and neighbouring countries. These findings suggest that double cropping in the NCP can amplify the magnitude of summertime climate changes over East Asia.