Paddy Water Environ (2010) 8:217–226 DOI 10.1007/s10333-010-0201-y

ARTICLE

The role of paddy rice in recharging urban groundwater in the Shira River Basin Kenji Tanaka • Yoshitaka Funakoshi • Takaomi Hokamura • Fumihiko Yamada

Received: 3 August 2009 / Revised: 15 September 2009 / Accepted: 18 February 2010 / Published online: 7 March 2010 Ó Springer-Verlag 2010

Abstract Agricultural fields in the middle Shira River basin play an important role as a source of groundwater recharge; however, the water balance between the agricultural water and river water is unclear. This study was conducted to investigate the water balance in the fields by measuring the stream flow of agricultural water channels, which draw water from the Shira River. The flow rate of water channels was found to increase in the beginning of May, which corresponded to the cultivation of paddy rice fields. During summer, the total agricultural intake was comparable to the river flow observed in the middle Shira River Basin. Determination of the water budget for the targeted area revealed that most of the recharged water was dependent on agricultural irrigation from the river. The annual recharge of the overall target area was estimated to be as high as 15,300 mm. In addition, the infiltration rate was as high as 170 mm/day in the paddy fields during summer, and as high as 30 mm/day in the upland fields during winter. In order to recover the groundwater recharge in this region, it is necessary to extend the submerged period to include periods in which the stream water in the Shira River is not subject to heavy rainfall as well. Keywords Kumamoto  Shira River  Agricultural water  Groundwater recharge  Water balance

K. Tanaka (&)  Y. Funakoshi  T. Hokamura  F. Yamada Department of Civil and Environmental Engineering, Graduate School of Science and Technology, Kumamoto University, Kumamoto, Japan e-mail: [email protected]

Introduction The Kumamoto Urban Region, which comprises the city of Kumamoto and its surrounding districts, is known as the largest urban groundwater region in Japan. Indeed, 100% of the water supply for nearly one million people in this region is dependent on groundwater. The quality of groundwater in Kumamoto is excellent, and it is renowned as the best-tasting water in Japan (Hashimoto 1989). However, a decline in the amount of water in the region has occurred over the last few decades in response to a decrease in the recharge area (Tsuru et al. 2006). Accordingly, local governments in the Kumamoto Urban Region have developed a plan designed to preserve the groundwater resources in Kumamoto (Kumamoto Prefecture 2009). These governments have estimated that the groundwater recharge in the region in 2007 was 6.00 9 108 m3/year, but that this will have decreased to 5.63 9 108 m3/year by 2024 if no countermeasures are implemented. Additionally, it has been determined that a recharge volume of 6.36 9 108 m3/year is necessary to preserve the groundwater, assuming that the groundwater pumping rate does not change from the rate that was observed in 2006 (1.862 9 108 m3/year). Hence, this project is designed to increase the amount of groundwater recharge in the region by 7.3 9 107 m3 by the year of 2024. The middle Shira River Basin is a key area for groundwater recharge. In this area, the geological structure of the surface layer has a high permeability. Specifically, the top several meters of surface soil consist of alluvium on a pyroclastic flow deposit known as Aso-III (approximately 132,000 years old) and Aso-IV (approximately 89,000 years old). Additionally, a large groundwater path extends from this area to the Kumamoto Plain, in which there is a

123

218

mean velocity of 40 m/day (Kumamoto Prefecture and Kumamoto City 2005). Field surveys have shown that the groundwater recharge area has a high capacity for recharge, with the soil infiltration rate ranging from 30 to 500 mm/ day (Kiriyama and Ichikawa 2004; Takemori and Ichikawa 2007). Based on these findings, it is believed that urbanization and increase in crops other than paddy rice have caused a decline in groundwater recharge. Although the importance of the Shira River middle area has been demonstrated, the surface water balance in the region is still unclear, especially the balance between river flow and agricultural water use. The effect of agricultural water use on river flow has been reported by several authors (Shimotsu 1987; Arai 2004), but there is a lack of quantitative evaluation based on the measurement of agricultural water channels. Because of this uncertainty, the water requirements for river flow in the Shira River have not yet been confirmed (MLIT 2002). Determination of this balance will enhance understanding of the capacity for groundwater recharge throughout the agricultural area. Many studies have been conducted to evaluate the water balance in the agricultural area, and these studies have revealed that the transfer of water from agricultural to urban use is an important issue (Revine et al. 2007; Matsuno et al. 2007; Huang et al. 2007). However, studies conducted to date have determined the quantitative balance using numerical models such as the Soil and Water Assessment Tool (SWAT) (Van Liew et al. 2003; Schilling et al. 2008), while information based on field measurements is very limited. Therefore, this study was conducted to investigate the water balance over the agricultural area and river flow in the middle Shira River Basin by measuring water flow in the irrigation channels. Specifically, the annual cycle and interannual variation in stream water over the middle Shira River Basin was analyzed using an MLIT observation data set. This information was then used to estimate the areaaveraged groundwater recharge rate for one of the local areas, Sako. When evaluating the recharge rate, the effects of evaporation flux on the surface energy balance were estimated with consideration of the land use cover.

Materials and methods Study area Figure 1 shows the location of the station and agricultural water channel system in the middle Shira River Basin, which contains the towns of Ozu, and Kikuyo and the northeastern portion of the City of Kumamoto. The bold line indicates the watershed boundary. The watershed of the middle Shira River Basin is narrow; therefore, the river

123

Paddy Water Environ (2010) 8:217–226

Fig. 1 Location of intake weirs on the Shira River and of the agricultural water observation stations used in this study. Stations 1– 11 in this figure correspond to the first column in Table 1. Major waterways from each intake weir are shown as bold lines. The bold break line indicates the boundary of Shira River Basin. Jinnai Waterstream Station maintained by the MLIT, Mashiki Station maintained by the JMA-AMeDAS, and Kumamoto JMA Observatory are also shown

receives very little inflow from tributaries. Seven intake weirs are in place for agricultural use in the study area: Hata-ide, Uwa-ide, Shimo-ide, Sako-Tamaoka, Tsukure, Babagusu, and Toroku. The hatched area shows the irrigation area of each agricultural waterway. Some agricultural water from the Shira River is returned to the river after use, while some is directed to other river systems including the Hori River from Uwa-ide and the Kase River from Toroku. In the Shira River, the Jinnai flow station operated by the MLIT (36.05 km from the mouth of the river) is located immediately downstream of the SakoTamaoka Weir. In this study, Sako was treated as the target area for the water budget, because both irrigation water (Sako 1, 2) and runoff water (Sako 3, 4) are measured in this area. Figure 2 shows the land use conditions in Sako determined during a field survey conducted 31 July 2008. The area of each land

Paddy Water Environ (2010) 8:217–226

219

Fig. 2 Land use in the Sako area in July 2008. The dotted line shows the area boundary

use type was calculated using GIS (Kumamoto GP Map) and is summarized in Table 1. In the upland fields, soybeans are primarily cultivated during summer. The land use category of the irrigation without planting represents the area that is seasonally submerged only for groundwater recharge. Paddy rice fields cover about 33%, average conditions during mid-summer (August 8) and low water conditions following the harvest of paddy rice (October 21, November 19–29).These data were then used in the following regression function (Tamura et al. 2006): pffiffiffiffi Q ¼ aH þ b ð1Þ where the values of a and b are those shown in Table 2. For the Sako 4 station, the flow rate was obtained from Manning’s equation (Yen and Tsai 2001): 1 Q ¼ I 1=2 R2=3 Ac n

ð2Þ

where n is Manning’s roughness coefficient (0.02), I the slope coefficient (1/345), R the hydraulic radius, and Ac a cross section of the open channel. In order to obtain the water level from the pressure gauge, it was necessary to correct the atmospheric pressure. The atmospheric pressure for each site, Pa, was calculated as follows (Kondo 1994):

Table 1 Land use condition in the Sako area in July 2008 Land use

Covered area (ha)

Paddy rice field Upland field

33.17 32.07

Irrigation without planting

7.31

Residences and buildings

10.17

Others (Roads, etc.)

19.34

Total

102.06

 Pa ¼ Pref

Ts Tref

g=CRd ð3Þ

where Pref is the reference atmospheric pressure at Babagusu, Tref the reference atmospheric temperature in Kelvin, Ts the atmospheric temperature at the target station, C the environmental temperature lapse rate (=6.5 K km-1), Rd the gas constant (287 J kg-1 K-1), and g the acceleration of the gravity. Atmospheric temperature can be calculated using the lapse rate, which is given by T = T0 ? C(z - z0), where T0 and z0 (=37.0 m) represent the atmospheric temperature and altitude at Kumamoto station. The corrected water level is nearly 10 cm higher at Hata-ide 1 station with the correction than without the correction. River flow measurement data set River flow data for Japanese national rivers (class 1) from recent years is available from the Water Information System (WIS) (URL http://www1.river.go.jp) maintained by the MLIT. The authors obtained hourly data regarding water levels and river flow observed at Jinnai Station (Fig. 1) from the WIS for Jan. 2004 to Apr. 2009. Because of the strict data quality check by MLIT, it takes several years for the flow rate to be released. In that period, the flow rate data were only released for 2004 and 2006. Hence, we calculated the flow rate for other periods (2005, 2007–2009) based on the hourly water level using the relationship described in Eq. 1, in which coefficients a = 5.4015 and b = 5.2609 were derived from flow rates and water levels observed during 2006. After determining the hourly flow rate, the daily average value was calculated. It should be noted that water level data for later than July 2008 are preliminary values. In addition to the aforementioned hourly data set, daily flow rate data from 1979 to 2003 were obtained from the

123

220

Paddy Water Environ (2010) 8:217–226

Table 2 List of stations at which flow rate was measured No.

Station Name

Latitude

1

Hata-ide 1

32°520 24.500 N

Longitude

Altitude (m)

a

130°570 10.800 E

b

181

1.3191

0.4001

2

Hata-ide 2

32°52 04.0 N

130°560 04.100 E

164

2.1784

-0.4416

3

Uwa-ide

32°520 17.400 N

130°550 59.100 E

145

4.8548

0.5186

0

0

0

00

00

00

4

Shimo-ide

32°52 12.5 N

130°55 40.9 E

143

2.4339

-0.9868

5

Sako-1

32°510 37.300 N

130°530 26.700 E

95.5

1.0113

0.1097

6

Sako-2

32°510 37.700 N

130°530 26,800 E

95.5

1.4711

0.0707

7

Sako-3

32°510 16.700 N

130°520 08.700 E

83.5

4.2112

0.1869

8

Sako-4

32°510 07.800 N

130°520 03.100 E

88.2

N/A

N/A

9

Tsukure

32°510 18.000 N

130°510 32.800 E

77.2

2.9338

-0.1045

0

0

00

00

10

Babagusu

32°51 07.7 N

130°50 56.6 E

65.7

2.7112

-0.177

11

Toroku

32°480 44.600 N

130°440 06.800 E

38.5

1.8274

0.1407

The number in the first column shows the station number plotted in Fig. 1. The right two columns indicate the regression coefficient in Eq. 1

rain and river flow database provided by the Japan River Association. These data were then used to investigate the interannual characteristics of river flow over the last 30 years. We first separated the 30-year data set by date-ofyear (DOY), after which we computed the 30-year average for each DOY. Statistical values such as the median, 25% lower flow and 10% lower flow were also extracted for each DOY by numerical sorting. Water budget for the targeted area The basic balance of the surface water budget integrated over the targeted area, A, can be given as:  Z  Z   X dg ðQirr  Qroff Þ þ Pr  EL  Rc dA þ  dA dt A

A

¼ 0; ð4Þ where Qirr irrigation from the river, Qroff surface runoff, Pr precipitation, EL net evapotranspiration from each land cover, Rc recharge into the ground, and g temporal water storage by the surface (e.g., inside the paddy fields). The first term on the left-hand side represents the water budget between the river basin and inside the agricultural field. The second term represents the vertical budget term and includes precipitation, evapotranspiration, and groundwater recharge. The third term, temporal storage, might be significant in short time scales (e.g., days), but is negligible on an annual scale. For the Sako Area, the first term on the left-hand side of Eq. 4 was determined based on field observations. Ground precipitation data are available from the Mashiki station (32°50.2N, 130°51.3E, 193 m a.s.l.) and AMeDAS, which is 2.5 km southwest of the Sako Area. However, the ground observations from 1600 JST 16 August to 1200 JST

123

17 August were missing; therefore, the precipitation intensity obtained from the 1-km Grid Point Value of the JMA Radar was inserted. The amount was estimated to be 70 mm during the missing period of ground observation. Evapotranspiration from the area was estimated using the monthly meteorological data obtained from Kumamoto Observatory, JMA (32°48.8N, 130°51.3E, a.s.l.; see Fig. 1b) because no observations of humidity or radiation were available at Mashiki Station and no micrometeorological observations were available for the target area. Estimation of evapotranspiration for each land use Evapotranspiration from the surface was computed based on the surface energy budget using the following equation (Arai 2004; Kondo 1994): E¼

Rn  G Lð1 þ BÞ

ð5Þ

where E the evapotranspiration flux, Rn the net radiation flux, G the soil heat flux, L the latent heat of water, and B = H/LE is the Bowen ratio. Net radiation flux can be computed as follows (Arai 2004):   pffiffiffiffiffi Rn ¼ ð1  aÞS  1  cLW n2 ð1  aLW  bLW ev ÞrT 4 ; ð6Þ where a the surface albedo, S the incoming shortwave radiation, r = 5.67 9 10-8 Wm-2 K-4 the Stefan–Boltzmann constant, T the atmospheric temperature in Kelvin, n the cloud cover ratio, ev the water vapor pressure (hPa), aLW = 0.51, bLW = 0.0061, and cLW = 0.64 (Arai 2004). The observed daily variables, S, n, ev and T, were obtained from the Kumamoto Observatory (Fig. 3). The difference in the monthly mean temperature between the observatory and the station (0.3–0.4 K) was about 0.01–0.02 MJ/day,

Paddy Water Environ (2010) 8:217–226

221

(a) Incoming solar radiation (MJ/m2day) 35 30 25 20 15 10 5 0 Jan−08

Table 3 Surface albedo and Bowen ratio for land use as determined by Inoue (2008) and Kondo (1994) Land use Paddy field Upland field

Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

0.08–0.20

0.13

(Summer)

0.14

0.29 0.27

Irrigation without planting

(Summer)

0.06–0.09

0.24

Residences and buildings

(Summer)

0.20

0.79

Others

(Summer) (Winter)

1.70 0.25

0.79 1.70

Results and discussion Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

Annual cycle of agricultural intake

Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

(d) Fractional cloud cover 1.0 0.8 0.6 0.4 0.2 0.0 Jan−08

(Summer)

(Winter)

(c) Vapor pressure (hPa) 35 30 25 20 15 10 5 0 Jan−08

Bowen Ratio

(Winter)

(b) Atmospheric temperature (degC) 35 30 25 20 15 10 5 0 Jan−08

Albedo

Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

Date Fig. 3 Meteorological variables observed at Kumamoto Observatory: a incoming solar radiation (MJ/day), b daily mean temperature (°C), c vapor pressure (hPa) and d normalized cloud cover ratio (0–1)

which is negligible when compared with the shortwave radiation. Soil heat flux was estimated using the thermal diffusion equation with the daily mean atmospheric temperature as a surface boundary condition. The soil heat flux was found to range from 7 to 10 Wm-2 during the warm season and from -5 to -12 Wm-2 during the cold season. In order to determine the Bowen ratio and the surface albedo, we used the typical value of each land use condition listed in Table 3 (Inoue 2008; Kondo 1994). These values were used because there are no recorded measurements for the atmospheric surface layer over the paddy fields in the middle Shira River Basin.

Figure 4 shows the variation in the daily average flow rates observed at each agricultural site. The annual cycle of agricultural water flow can be seen clearly, except for the Babagusu water channels. The irrigation rate into the agricultural area increases in the beginning of May, approximately 1 month before the paddies are sown (Fig. 5). The inflow rate for each agricultural channel fluctuates during the rainy season due to control of the inner flood water and is higher from July to late September than during the rest of the year. During this period, the irrigation rate ranges from 6 to 9 m3/s at Uwa-ide and Shimo-ide, from 4 to 5 m3/s at Babagusu and Toroku, and from 1.5 to 2.5 m3/s at Sako and Tsukure, respectively. The flow rate decreases rapidly in the beginning of October, which corresponds to the harvest period of rice. The irrigation rate is maintained at a low level from late October until the following April, during which time wheat is grown in the paddy fields. Based on the results shown in Fig. 4, the water balance can be separated into three phases: increased irrigation in May and June (Phase-I), high irrigation from July to early October (Phase-II), and no rice cropping from mid October to the following April (Phase-III). Figure 6 shows a flow diagram of the middle Shira River averaged for each phase. The total intake water above Jinnai Station (Hata-ide 1, Uwa-ide, Shimo-ide and Sako-1, 2) was 10.38 m3/s during Phase-I, 15.27 m3/s during Phase-II, and 4.29 m3/s during Phase-III. Additionally, the total intake below Jinnai Station (Tsukure, Babagusu and Toroku) was 5.93 m3/s during Phase-I, 9.50 m3/s during Phase-II, and 3.03 m3/s during Phase-III. The total intake from Shira River was 16.31 m3/s (8.611 9 107m3 for amounts), 24.77 m3/s (2.183 9 108m3), and 7.32 m3/s (1.277 9 108 m3) for each phase, which was 35.2, 90.6, and 32.8% of the river flow at Jinnai

123

222

Paddy Water Environ (2010) 8:217–226

(a) Upper Middle Shira River

Flow rate (m3/s)

Hata−ide 1

10 9 8 7 6 5 4 3 2 1 0 Jan−08

Apr−08

Hata−ide 2

Jul−08

Uwa−ide

Oct−08

Shimo−ide

Jan−09

Apr−09

(b) Sako Area Inflow (Sako−1,2)

Runoff (Sako−3,4)

Flow rate (m3/s)

3.0 2.5 2.0

Fig. 6 Flow diagram of the middle Shira River basin separated by cropping phase: beginning of the rice-cropping season (from May to June 2008), middle of the rice-cropping season (from July to 10 October 2008) and off-season cropping (from 11 October 2008 to April 2009)

1.5 1.0 0.5 0.0 Jan−08

Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

(c) Lower Middle Shira River Basin

Flow rate (m3/s)

Tsukure

10 9 8 7 6 5 4 3 2 1 0 Jan−08

Apr−08

Babagusu

Jul−08

Oct−08

Toroku

Jan−09

Apr−09

Fig. 4 Variation in the daily average flow rate observed at each station. a Hata-ide 1,2, Uwa-ide, Shimo-ide, b Sako 1–4 and c Tsukure, Babagusu, Toroku

Station, respectively. The annual amount of intake water was 2.643 9 108 m3 on the upstream side of Jinnai and 1.678 9 108 m3 on the downstream side, while the annual river flow at Jinnai was 8.777 9 108 m3. It should be noted that the river flow rates were sensitive to rainfall (Fig. 3). The hydrograph of the middle Shira River is shown in Fig. 7. The bars show the daily Fig. 5 Standard water level control over the paddy field according to the cropping cycle in the middle Shira River Basin. The original agricultural calendar is provided by the Kikuchi branch office of Japan Agriculture (JA)

123

precipitation observed at Mashiki Station. High water days in June and early October, when the flow rate was greater than 100 m3/s, were found to correspond to rainfall levels of greater than 60 mm/day. The stream water flow rate under the high water conditions (9 days) in 2008 was 1.425 9 108 m3, or about 16.3% of the annual amount of stream water (32.2% of the amount during the rice-cropping period). Except for these high water days, the river streamflow ranged from 15 to 30 m3/s. The streamflow decreased between late April and early May in 2008, with the minimum streamflow being observed in the beginning of June. The statistical properties of river flow in the middle Shira River are plotted in Fig. 8. The interannual variation in river flow can be clearly seen from May to October. The maximum flow rate for each DOY depends on the individual mesoscale systems, such as the front system of the extratropical cyclone and typhoon. The median flow rate ranged from 15 to 20 m3/s from August to the following April. The median flow decreased to less than 10 m2/s in the beginning of June and then increased to 40 m3/s during

Paddy Water Environ (2010) 8:217–226

223

Flow rate (m3/s)

Flow rate at Jinnai

Daily precipitation at Mashiki 0

600 550 500 450 400 350 300 250 200 150 100 50 0

40 80 120 160 200 240 280 320

Precipitation (mm/day)

Fig. 7 Hydrograph of the middle Shira River Basin. The flow rate observed by Jinnai Station and daily rainfall at Mashiki are shown

360 400 Jan−08

Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

Date

Statistical properties (Jinnai) (1979−2008) Maximum Average 25% Low

Minimum Median 10% Low

Flow rate (m 3/s)

1000 500 200 100 50 20 10 5 2 1 0.5 0.2 0.1 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Date Fig. 8 Statistical properties of river flow observed at Jinnai in the middle Shira River plotted by date

the rainy season of June and July. The 25% lower flow and 10% lower flow falls below 10 m3/s during the rice-cropping season, while these lower flow values are greater than 10 m3/s during the off-season. Such depression of the lower flow values implies that the water intake is too high during dry periods. Therefore, further studies should be conducted to evaluate the interannual variation in the water balance between agricultural channels and the river. Water budget of the local area Table 4 summarizes the water budget of the targeted area from May 2008 to April 2008. Each component of the water budget in Eq. 4 was translated into millimeters (Table 4). The annual precipitation at Mashiki Station was 2463 mm in 2008, which was comparable to the

precipitation observed during the study period (2493 mm). The water level in the paddy fields only varied by several centimeters (Fig. 5), which was much smaller than the variation in net irrigation that was observed over the entire study area. The average annual groundwater recharge for the targeted area was estimated to be 15,290 mm. The annual evapotranspiration was estimated to be 657.9 mm, which was approximately 15% lower than in previous studies. For example, Suekane and Kayane (1980) estimated the annual evapotranspiration to be 777 mm using Penmann’s method. The estimated evapotranspiration of July and August was nearly the same as that estimated by Suekane and Kayane (1980); however, in winter and spring, the results of the present study were 10–20 mm/ month lower than in previous studies. Since Penmann’s method is for homogeneous open-wet surfaces, it tends to overestimate the heterogeneous land cover; therefore, it must be corrected using a reduction factor obtained by field observations (Yabusaki et al. 2005). The daily evapotranspiration from each land use condition is shown in Fig. 9. Under fine weather conditions, the daily evapotranspiration was higher than 6 mm/day from paddy fields that were irrigated without planting during summer, whereas it was as low as 3 mm/day for residential areas. The average daily evapotranspiration for the entire study area was about 5 mm/day under fine weather conditions during summer. For all land use conditions, the daily variation in evapotranspiration fluctuated with the incoming solar radiation as shown in Fig. 3a. Under cloudy and rainy conditions, the evapotranspiration was smaller than 1 mm/ day, even in June. Considering the complex land cover in the targeted area, the present estimation is reasonable. The recharge rate Rc was estimated assuming that the storage term was negligible. The dependency in Table 4 represents the ratio of net irrigation to the residual recharge Rc. The recharge rate during the rice-cropping period

123

224

Paddy Water Environ (2010) 8:217–226

Table 4 Monthly water budget for the Sako area from May 2008 to April 2009 Qin–Qout (mm)

Month

P (mm)

E (mm)

P–E (mm)

Rc (mm)

Dependency (%)

May 2008

1412.2

227.5

77.7

149.8

1562.0

90.4

Jun. 2008

1910.1

764.5

71.1

693.4

2603.5

73.4

Jul. 2008

2150.1

229.0

133.4

95.6

2245.8

95.7

Aug. 2008

2325.6

177.5

104.2

73.3

2398.9

96.9

Sep. 2008

1977.7

331.0

69.0

262.0

2239.7

88.3

Oct. 2008

655.8

88.5

50.7

37.8

693.6

94.5

Nov. 2008

593.5

119.0

19.1

99.9

693.3

85.6

Dec. 2008

482.0

114.0

4.9

109.1

591.2

81.5

Jan. 2009

680.7

57.5

11.4

46.1

726.8

93.7

Feb. 2009

529.5

152.0

16.5

135.5

665.0

79.6

Mar. 2009 Apr. 2009

560.7 177.8

149.0 83.5

39.4 60.6

109.6 22.9

670.3 200.8

83.6 88.6

13455.8

2493.0

657.9

1835.1

15290.9

88.0

Total

(2200–2600 mm) was approximately three or four times higher than that of the no rice-cropping period (600– 700 mm). Most of the recharge water depends on the irrigation water, and the recharge rate ranges from 80 to 97% throughout the year. Here, we attempt to estimate the daily recharge rate over the paddy fields and temporal changes in the water surface during irrigation. In order to accomplish this, the following equations were used: Reff ¼

Rcr Aall  Wl A p þ Aw

Wl ¼ Rcn Aall

Au Agr

ð7Þ ð8Þ

where Rcr the Rc during the rice-cropping period, which was determined to be 76.1 mm/day based on the information in Table 4, Rcn the Rc during the off-season cropping period, which was determined to be 22.2 mm/day based on the average from Nov. 2008 to Mar. 2008, Aall the total area of the targeted area, Ap the area of the paddy fields, Aw the area of irrigation without planting during summer, Au the area of the upland fields during summer, and Agr = Ap ? Aw ? Au the total agricultural field area. Wl in Eq. 8 represents the infiltration water mass over the upland fields shown in Fig. 2 assuming that the recharge rate is constant throughout the year. Reff was estimated to be 167.5 mm/day. Takemori and Ichikawa (2007) demonstrated that the infiltration rate of paddy fields in this area was 130–300 mm/day. Hence, the estimation obtained in this study is reasonable. In addition, the daily recharge rate during winter was estimated to be 31.3 mm/day. Taken together, these results indicate that the annual total recharge in the paddy fields is 27,740 mm/year (120 days for paddy rice irrigation), while it is 11,380 mm/year in the upland fields.

123

Figure 10 shows the variation in area coverage of the paddy rice field and the total submerged area (paddy field and irrigation without planting) for the towns of Ozu and Kikuyo in the middle Shira River Basin. The acreage of the paddy rice fields in this area was as high as 1600 ha during the 1970s, which resulted in a contribution to groundwater recharge of about 3.0 9 108 m3; however, this value has decreased gradually since the 1980s. The acreage of paddy rice has been lower than 1000 ha since 2000, and in recent years the acreage was half that observed in the 1970s (i.e., about 1.5 9 108 m3). The reduction of paddy fields in the middle Shira River over the last few decades has affected the reduction of groundwater by about 1.5 9 108 m3, which is about 25% of the annual recharge in 2008 estimated by Kumamoto Prefecture. Since 2004, the seasonal flooding campaign has been extended to fallow fields during summer to increase the groundwater recharge. Specifically, 291 ha were flooded in 2004, while 576 ha were flooded in 2008. It is estimated that an additional 1.637 9 107 m3 of water was recharged by this campaign in 2008, assuming 30 days of submersion with a recharge rate of 100 mm/day (Kumamoto Prefecture 2009). The areal extension of the submerged surface for groundwater recharge in the agricultural area will compensate the required water to some extent, but the period should be made as long as possible in accordance with the agricultural cycle. In order to increase the groundwater recharge over the agricultural area, the intake from the Shira River must be increased by at least 5 m3/s during the rice-cropping period. However, increasing the intake prior to the rainy season could lead to critically low water levels, as shown in Figs. 6 and 8. Furthermore, the irrigation gates are controlled to shut off under high water conditions such occur

Paddy Water Environ (2010) 8:217–226

225

Area Avg.

Others

Residence

Irr. w.o. Plant.

2000

Acreage (ha)

Upland field

Paddy field

Daily Evapotranspiration (mm) 10 8 6 4 2 0 10 8 6 4 2 0 10 8 6 4 2 0 10 8 6 4 2 0 10 8 6 4 2 0 10 8 6 4 2 0 Jan−08

Total Seasonally flooded area Paddy rice in Ozu Paddy rice in Kikuyo

1500

1000

500

0 1960

1970

1980

1990

2000

2010

Year Fig. 10 Interannual variation of rice-paddy acreage of the towns of Ozu and Kikuyo in the middle Shira River Basin, provided by the Kyushu Regional Agricultural Administration Office. The area of seasonal flooding is based on data corresponding to Kumamoto Prefecture (2009)

Apr−08

Jul−08

Oct−08

Jan−09

Apr−09

Date Fig. 9 Daily evapotranspiration from each land use condition as listed in Table 2

during the Baiu season. For example, in 2008, 1/3 of the stream water was not available during summer. An increase in heavy rainfall events as a result of recent climate change may lead to further reductions in the available stream water during summer. Hence, it is more efficient to recharge the groundwater during winter using idle fields than by increasing the irrigation area during summer.

Concluding remarks This study investigated the water balance between agricultural water and river flow by measuring water flow in water channels in the Shira River Basin. The annual cycle of water flow into the agricultural area can be clearly seen for each water channel except

Babagusu, in which the control of the gate is more complex. The agricultural intake of water has significant impacts on river water in the middle Shira River Basin. The period of no-precipitation days during the paddy ricecropping season led to a severe reduction in the main streamflow. The total agricultural intake during summer was as high as the flow observed in the middle of the Shira River. Even during winter, the amount of agricultural water collected was about 30% of the water flow in the main river. The annual amount of intake water was 2.643 9 108 m3 on the upstream side of Jinnai and 1.678 9 108 m3 on the downstream side, while the annual river flow at Jinnai was 8.777 9 108 m3. Development of a water budget of the targeted area revealed that most of the recharged water was dependent on agricultural irrigation from the Shira River. The daily recharge rate was *166.9 mm/day in the paddy rice fields during summer and about 30 mm/day in the upland fields during winter. The annual recharge over the targeted area was 15,000 mm, while it was 27,800 mm/year in the paddy fields. The current recharge in the paddy field in the middle Shira River Basin was 1.5 9 108 m3, which was about half of the intake water from the Shira River reported for 2008. The reduction of the cropping area in the middle Shira River led to a reduction in groundwater recharge of 1.5 9 108 m3 when compared with the level of recharge in the 1970s. Such a reduction is as high as 25% of the annual recharge estimated by Kumamoto Prefecture for 2008. In order to increase the groundwater recharge to preserve water resources in the Kumamoto Urban Region, it is difficult to depend only on the middle Shira River. Most of the water that recharges the agricultural area originates

123

226

from surface runoff in the upper basin (e.g., from inside the caldera of Mt. Aso). If the recharge over this area must be increased, it would be more efficient to irrigate idle fields during winter because the river flow in winter is much more stable than in summer. It should be noted that Tomiie et al. (2009) found that the concentration of Nitrate–Nitrogen (NO3–N) in groundwater has increased gradually throughout the Kumamoto Urban Region, especially in northeastern Kumamoto. However, the primary cause of such contamination is still not clear. Therefore, the effects of irrigation on water quality should be addressed in future studies. Acknowledgments This study was supported by a grant from the Center for Politics Study, Kumamoto University. The authors also thank Mr. Hirano of the Sea-Bass Planning Co. Ltd. for providing the flow measurement data for the water channels.

References Arai T (2004) Hydrology for regional analysis. Kokon Shoin Publishers, Tokyo, 309 pp Hashimoto S (1989) Evaluation of water quality and/or tasty drinking water and its application to Japanese waters. J Soc Heat AirCond Sanit Eng Jpn 63:463–468 Huang CC, Tsai MH, Lin WT, Ho YF, Tan CH, Sung YL (2007) Experiences of water transfer from the agricultural to the nonagricultural sector in Taiwan. Paddy Water Environ 5:271–277 Inoue K (2008) Evaluation of Bowen ratio and the climate-moderating index on agricultural land in clear summer day. J Agric Meteorol 64:157–166 Kiriyama T, Ichikawa T (2004) Preservation of ground water basin recharging by paddy field. Annu J Hydraul Eng JSCE 48: 373–378 Kondo J (ed) (1994) Hydrometeorology. Asakura Shoten, Tokyo, 368 pp Kumamoto Prefecture (2009) 1st Implementation plan for groundwater resource management, http://www.pref.kumamoto.jp/ soshiki/48/dai1ki-koudoukeikaku.html (in Japanese). Last updated on 23 Feb 2009, last viewed on 28 Sep 2009

123

Paddy Water Environ (2010) 8:217–226 Kumamoto Prefecture and Kumamoto City (2005) The investigation report on the mechanism of groundwater contamination by nitrate, Kumamoto, Japan Matsuno Y, Hatcho N, Shindo S (2007) Water transfer from agriculture to urban domestic users: a case study of the Tone River Basin, Japan. Paddy Water Environ 5:239–246 Ministry of Land Infrastructure and Transportation (MLIT) (2002) River development project for Shira-River system, 61 pp Revine G, Barker R, Huang CC (2007) Water transfer from agriculture to urban uses: lesson learned, with policy considerations. Paddy Water Environ 5:213–222 Schilling KE, Jha MK, Zhang Y-K, Gassman PW, Wolter CF (2008) Impact of land use and land cover change on the water balance of a large agricultural watershed: historical effects and future directions. Water Resour Res 44 W00A09. doi: 10.1029/2007WR006644 Shimotsu M (1987) Groundwater circulation system over west outer rim of Mt Aso. J Jpn Assoc Groundw Hydrol 29:161–170 Suekane A, Kayane I (1980) Response of groundwater table to rainfall in Kumamoto Plain. Geogr Rev Jpn 53:666–671 Takemori Y, Ichikawa T (2007) On effect of groundwater evaluation by keeping water in no use paddy field in the middle Shira-River Area, vol 34. Bulletin of School of Engineering, Kyushu Tokai University, Japan, pp 1–8 Tamura T, Hashino M, Tachibana D (2006) Parameter identification of runoff model using rainfall and water-level data. Annu J Hydraul Eng JSCE 50:355–360 Tomiie K, Iwasa Y, Maeda K, Otuzuki M, Yunoue T, Kakimoto R, Kawagoshi Y (2009) Present status and feature of groundwater contamination by nitrate-nitrogen in Kumamoto City. J Water Environ Technol 7:19–29 Tsuru N, Akiyoshi H, Miyamoto H (2006) Groundwater pollution by nitrate-nitrogen in Kumamoto city, vol 14. Annual Report of Kumamoto City Environmental Research Institute, pp 59–67 Van Liew MW, Schneider JM, Garbrecht JD (2003) Streamflow response of an agricultural watershed to seasonal changes in precipitation. In: Proceedings of the 1st interagency conference on research in the watersheds (ICRW), Oct. 27–30, 2003, United States Department of Agriculture (USDA) Yabusaki S, Tase N, Haginoya S (2005) Consideration for the estimation methods of evapotranspiration at Terrestrial Environment Research Center, vol 6. Bulletin of the Terrestrial Environmental Research Center (TERC) University of Tsukuba, Japan, pp 45–51. Yen BC, Tsai CW-S (2001) On noninertia wave versus diffusion wave in flood routing. J Hydrol 246:97–104