Effects of land use change on hydrological cycle from forest to upland field in a catchment, Japan

Abstract

Understanding the effects of land use change on the hydrological cycle is very important for development of sustainable water resource in an upland field catchment. In this study, soil and hydrological properties in an upland field catchment, which was reclaimed partially from a forest catchment, were compared with another forest catchment. The soil properties of surface and subsurface layers were investigated in the two catchments. The soil was compacted and water-holding capacity of soil in the upland field catchment became smaller after the reclamation from forest to upland field, which decreased infiltration rate and water storage in the soil layers. We found that peak discharge and direct runoff in the upland field catchment increased compared with the forest catchment. Annual evapotranspiration from the upland field catchment tended to be lower due to the change in vegetation type and soil properties. Furthermore, a semi-distributed hydrological model was applied in the upland field catchment to understand the integrated effects of reclamation on the hydrological cycle. The model parameters, which were determined using a nonlinear optimization technique—the Shuffled Complex Evolution method (SCE), were compared between the two catchments. The Nash and Sutcliffe coefficient was used to evaluate the model performance. The simulated results indicated that evapotranspiration was decreased and change in discharge was more obvious in the surface layer. We considered that declined infiltration and water storage and increased peak discharge and direct runoff have a negative impact on water resources in the upland field catchment. This study will provide information for forest managers in planning and making decisions for land and water resource management.

Keywords

land-use change / hydrological processes / upland field catchment / forest catchment / semi-distributed hydrological model

Introduction

Land use is an important factor influencing hydrological processes. Changing land use can influence flood frequency [1-5], annual discharge [6], and evapotranspiration [7]. Evaluating the effects of land use changes on hydrology are the basis for catchment management and ecological restoration [8-10]. LØrup et al. [11] assessed the effect of land use change on six medium-sized rural catchments in Zimbabwe. They indicated that annual runoff decreased for most of the six catchments, with the largest changes occurring in catchments located within communal land. Niehoff et al. [12] analyzed the relationship between land-use change and hydrological response in a meso-scale catchment in south-west Germany. Choi and Deal [13] indicated that land-use changes resulted in little change in total runoff but noticeable change in surface flow.

In Japan, many national projects of reclamation of upland fields had been established after the Second World War in order to expand farmland area and to stabilize the farmer’s economy. However, some problems about abandoned farmland have become more serious recently. Now, it is important to know the effects of land use change on the hydrological cycle in a watershed for sustainable development and ecological restoration. Takase and Sato [14] compared evapotranspiration in an upland field catchment and forest catchment, and concluded that evapotranspiration in the upland field catchment was more restrained in summer compared with in winter, while it was not so different in the forest catchment. Hong et al. [15] indicated that runoff coefficient of peak discharge in an upland field catchment was slightly larger and the concentration time was shorter than those in a forest catchment. Wang et al. [4,16,17] analyzed storm-induced landslides using a hydrological model in a forest catchment; they concluded that rainfall intensity and the level of upper layer storage reflects cohesion and pore water pressure, and are the critical factors that trigger a land-slide. However, integrated effect on each hydrological cycle due to reclamation of upland field catchment is still unknown.

The objective of this study is to assess the changes in soil and hydrological properties in an upland field catchment after reclamation from a forest. Three aspects, which are investigation of soil properties, comparison of hydrological processes and application of hydrological model, were analyzed to understand effects of reclamation on hydrological cycle in the upland field catchment. Due to the difficulty of measurement of runoff from a forest area in the upland catchment, another neighboring forest catchment was selected as a reference for determination of the model parameters of the forest area and comparison of a water balance with the upland field catchment. To understand the topographical condition and soil properties such as bulk density and water-holding capacity were investigated after reclamation in the upland field catchment, and the evapotranspiration rate, peak discharge and direct runoff rate were compared with those of the forest catchment. To analyze integrated effects of the direct changes in vegetation and soil properties on hydrology, a semi-distributed hydrological model, comprising two types of land use (forest area and upland field area), was applied to simulate discharge and infiltration rate in the upland field catchment.

Study area

In this study, two experimental catchments were selected which are shown in Fig. 1. A forest catchment was selected as a reference catchment against a upland field catchment. It is located at 132°37′ E and 33°28′ N in Ehime Prefecture, Japan and includes broad leaf trees and conifer trees (Japanese cedar trees and Japanese cypress trees). The total area of the forest catchment is 0.21 km² and 60% of the area is covered by broad leaf trees. The elevation ranges from 246 m to 420 m. Discharge and rainfall were measured from August, 1984 to 2012 using a triangular weir, which was constructed on bed rock, and rainfall gauge, respectively.

The upland field catchment is located at 132°30′ E and 33°30′ N. It includes upland fields reclaimed in 1979 by a national project. The total area of the catchment is about 0.12 km², 58% of which was reclaimed as upland field and the elevation ranges from 25 m to 115 m. In most of the fields, watermelon and Chinese cabbage are grown in summer and winter, respectively. The forest in the upland field catchment contains mainly Japanese cypress trees and pine trees. Discharge and rainfall were measured from April 1979 until August 1995 using a rectangular weir and rainfall gauge, respectively.

Change in soil properties

An upland field in an experimental catchment was reclaimed by cutting ridge, filling valley and constructing roads and drainage channels. Therefore, dramatic changes were found in topographical features. The slope of the field area was 6 degree while original forest was 27 degree. The stream density was increased from 70 to 250 m/ha due to establishment of the drainage system.

Table 1 shows some changes in soil properties after reclamation in the upland field catchment. We found that soil dry density had been much increased in surface (5–10 cm) and subsurface soil layer (30 cm) just after reclamation. But that of surface soil was reduced for several years.

The relationship between water-holding capacity and volumetric water content (soil moisture characteristic curve) was also compared. Figure 2 shows three curves at original condition (circle solid line), immediately just after reclamation (asterisk dash line) and 7 years after reclamation (triangle dash line). From this figure, we concluded that the soil water-holding capacity was decreased due to the compaction of the soil and a great change was found in the range from saturation to pF2.0, which means that large pores were compacted primarily by the reclamation. However, the capacity had been recovered by farming activities such as tillage and cultivation.

Change in hydrological properties

Evapotranspiration

Evapotranspiration is an important factor of hydrological cycle. It is not only an inevitable process for all of vegetation but also a great loss of water resources. Furthermore, evapotranspiration plays an important role to redistribute the solar energy which is received on surface of the earth. Therefore, understanding the change in evapotranspiration process due to land use change is required if water sustainability is to be achieved. For a closed catchment, a simple annual water balance method was applied to estimate annual evapotranspiration (ET):

(1)ET=R-Q,

where ET is evapotranspiration; R is rainfall and Q is discharge.

To evaluate the change in evapotranspiration due to reclamation, we compared the ratios of the evapotranspiration to potential evaporation (ET/Ep) in the two catchments, and plotted the relationship with the annual rainfall (Fig. 3). Here, potential evaporation (Ep) was estimated by Penman’s equation using meteorological data which was collected from official meteorological station. The linear dependence of evapotranspiration ratio on annual rainfall reflects that evapotranspiration in the upland field catchment is less than that in the forest catchment. Thus, we considered that evapotranspiration from the upland field catchment may be influenced observably due to the small root zone of crops and the low available soil moisture for evapotranspiration.

Storm runoff and direct runoff

Estimating peak discharge in a catchment is one of the most important work for flood protection as well as of greatest overall economic importance. Figure 4 shows an example of storm hydrographs in the upland field and forest catchments. They show that the storm runoff in the upland field catchment was more sensitive to storm rainfall. In addition, the peak discharge and direct runoff rate in the upland field and forest catchments through five storms were compared, and the results are shown in Tables 2 and 3. Compared to the peak flow for the same rainfall intensity, the peak discharge from the upland field catchment increased 2.0-5.7 times compared with the forest catchment. Comparison of the direct runoff rates between the two catchments shows that the direct runoff rate is smaller in the forest catchment (from 20.0% to 46.1%) than in the upland field catchment (from 29.8% to 57.1%). The above results show that relatively more rainfall was saved in the forest catchment.

Hundreds of different methods have been used to estimate the peak discharge in drainage basins [18]. A traditional approach is a rational method that has been widely used around the world for flood estimation. It is generally considered to be an approximate deterministic model representing the flood peak that results from given rainfall, with the runoff coefficient being the ratio of the peak rate of runoff to total rainfall. Flood estimation is given by

(2)Q_p=\alpha \cdot r{e}_{tp}\cdot A,

where Q_p is a peak discharge; α is a unit conversion factor which is a constant; re_tp is average effective rainfall intensity during time tp; and A is catchment area. In this study, effective rainfall was calculated by re_tp = f·i, in which f is runoff coefficient, which is empirically defined as the ratio of peak discharge to average actual rainfall intensity during tp; and tp is concentration time of peak discharge. In applying the rational equation, Kadoya and Fukushima [19] proposed an empirical equation for the relationship between tp and re_tp;

(3)tp=C\cdot A^{\mathrm{0.22}}\cdot r{e}_{tp}^{-\mathrm{0.35}},

where C is a parameter dependent on land use. Figure 5 shows relationships between equivalent depth of peak discharge, which is equal to effective rainfall intensity, and an index of concentration time (tp/A^0.22). We found that the concentration time in the upland field catchment was smaller than that in the forest catchment. The values of C were 130 and 60 for the forest and upland field catchments, respectively.

Peak discharge was compared in an upland field and a forest under the conditions of same drainage area of 1 km², a return period's rainfall intensity of 50 mm/h and supposing that f are 0.5 for the upland field and 0.7 for the forest. A relationship between tp and rtp is given by the Sherman type equation. The calculation results showed that the peak discharge in the upland field was increased 2.5 times compared with the forest. We concluded that the reduction of infiltration and shortening of concentration time synergistically increased the peak discharge.

Development of a semi-distributed hydrological model

Outline of semi-distributed hydrological model

A semi-distributed model was applied in the upland field catchment to assess the effects of land use change on hydrological cycle between upland field and forest areas. The principal structure of model is almost same as the hydrological model which was proposed by Takeshita et al. [20]. The outline of model is shown in Fig. 6. It includes almost hydrological processes such as infiltration, percolation, runoff or evapotranspiration to represent the hydrological cycle in an upland field catchment.

Horton’s equation was introduced in the model as a physically-based infiltration sub-model:

(4)F(t)=FS+(FMAX-FS)\cdot \exp (-B\cdot t),

where F(t) is the infiltration rate at time t; FS is the final infiltration rate; FMAX is the maximum rate, corresponding to the infiltration rate when the surface soil layer is perfectly dry; B is a constant; and t is time. According to this equation, the maximum soil moisture, soil moisture at any time and soil moisture at pF1.8 are derived as following equations:

When F (t)>F18:

(5)S(t)={\displaystyle {\int }_{\mathrm{0}}^T(FS+(FMAX-FS)\cdot \exp (-B\cdot t))\mathrm{d}t.}

When F (t)<= F18:

(6)S(t)={\displaystyle {\int }_{T\mathrm{18}}^T((FMAX-FS)\cdot \exp (-B\cdot t))\mathrm{d}t},

(7)S\mathrm{18}={\displaystyle {\int }_{\mathrm{0}}^{T\mathrm{18}}(FS+(FMAX-FS)\cdot \exp (-B\cdot t))\mathrm{d}t,}

(8)S{M}_{\max }={\displaystyle {\int }_{\mathrm{0}}^{T\mathrm{18}}(FS+(FMAX-FS)\cdot \exp (-B\cdot t))\mathrm{d}t}+{\displaystyle {\int }_{T\mathrm{18}}^{\infty }(FMAX-FS)\cdot \exp (-B\cdot t)\mathrm{d}t},

where S(t) is the infiltration at any time; F18, and S18 is the rate and amount of soil moisture held at a gravity drainage of pF1.8 respectively, and SM_max is the maximum amount of soil moisture at a saturated condition.

In this hydrological model, we supposed that a ratio of evapotranspiration to potential evaporation (ET/Ep) was function of soil moisture and Eq. (9) was used to calculate the ratio for any amount of soil moisture which was calculated by Eq. (5) or (6)

(9){\displaystyle \frac{ET}{Ep}}=\left({\displaystyle \frac{\mathrm{2}}{\mathrm{1}+\exp \left(-AL\cdot {\displaystyle \frac{S(t)}{S\mathrm{18}-S(t)}}\right)}}-\mathrm{1}\right)\cdot EC{C}_{\max },

where ET is actual evapotranspiration; Ep is potential evaporation; AL is a model parameter; and ECC_max is a coefficient which represents a maximum ratio of evapotranspiration to potential evaporation.

Model calibration and validation

The model parameters were optimized using a mathematical optimization method—the Shuffled Complex Evaluation (SCE) method with observed discharge data from 1985 to 1989. First, the model parameters of the forest area were optimized using the data which were observed in the forest catchment. Then, the parameters of the upland field area were optimized using the data which were observed in the upland field catchment. In the optimization process, it was assumed that the parameter for groundwater layers was the same for the forest and upland field, because of no obvious effects in the deeper soil layer after reclamation in the upland field catchment.

The model parameters were optimized as to minimize the total sum of relative error between observed and calculated discharge given by following equation:

(10)Error={\displaystyle \sum _i^n{\displaystyle \frac{\left(q^{cal}-q^{obs}\right)}{q^{obs}}}}j,

where q^cal is the calculated discharge by the model; q^obs is the observed discharge.

Furthermore, the Nash and Sutcliffe coefficient [10,21,22] was used to evaluate the model performance, which was calculated using Eq. (11):

(11)R_{ns}=\mathrm{1}-{\displaystyle \frac{{\displaystyle \sum _i^n{w}_t^{\mathrm{2}}{(q_{}^{obs}-q_{}^{sim})}^{\mathrm{2}}}}{{\displaystyle \sum _i^n{w}_t^{\mathrm{2}}{\left(q_{}^{obs}-\overline{q_{}^{obs}}\right)}^{\mathrm{2}}}}}.

where q^obs is the observed discharge; q^sim is the simulated discharge; n is the number of discharge data; and w_t is weight (in this study, w_t = 1).

Results and discussions

The model parameters of a forest area were first optimized using the observed data from a forest experimental catchment. In the optimization process, the recession coefficient of the lowest tank was determined by inspection of observed hydrograph. Other parameters were optimized using the Shuffled Complex Evolution (SCE) method [23-25]. Figure 7 shows an example of the optimized result for the forest catchment. The mean relative error between the observed and calculated discharge in five years (1985-1989) was 18.0%, which indicates that the hydrological model can represent actual hydrological processes in the forest area of the upland field catchment [15]. The parameters of an upland field area were optimized after the calculation of discharge from the forest area in the distributed model.

The discharge in the upland field catchment in five years (1985-1989) was calculated. The calculation of discharge is matched very well with the observation discharge (Fig. 8). The mean relative error and model performance was 23.8% and 0.87, respectively. The discharge in the upland field catchment in 1995 was simulated to evaluate model validity. The model efficiency coefficient was 0.92 and relative error was 25.6%. As a result, the model structure and model parameters can reflect the hydrological processes in the upland field catchment based on the Nash and Sutcliffe coefficient criteria [3,4,26].

Values of optimized parameters for the forest area and upland field area are compared in Table 4. The final infiltration rate (FS) in the reclaimed area was less than that of forest area, which indicates that excess rainfall for storms in upland field catchment was increased. Moreover, integrated effects should be discussed because the other parameters, such as a maximum infiltration rate (FMAX) and recession parameter (B) also influenced the properties. Typical infiltration rates in both areas during a rainy season (1st – 3rd, September) are shown to discuss integrated effects (Fig. 9). It is obvious that the infiltration rate was decreased immediately in the upland field by rainfall and more excess rainfall produced in the upland field while no excess rainfall occurred in the forest.

For evapotranspiration process, parameters ECCmax and AL of the upland field area are less than those in the forest area. Figure 10 shows dependency of ratio of actual evapotranspiration to potential evaporation, which is formulated by Eq. (8). As shown in this figure, evapotranspiration in the upland field area decreased more than in the forest area when soil was dry. This result was similar to the result reported by Takase and Sato [14].

Comparison of runoff properties is difficult because not only runoff coefficient (Ci) but also percolation rate (GGi) influences runoff from soil layer. Therefore, annual water balance was compared by the simulation results for the forest and upland field areas. Table 5 shows a comparison of average water balance in the upland field catchment. Under a same hydrological condition, in which annual rainfall (R) and potential evaporation (Ep) were 1697.5 and 962.5 mm, respectively, actual evapotranspiration (Et) in upland field area was lower than that in forest area and total discharge (Qt) was greater in upland field area than in forest area. The discharge from the impermeable area (QCN) and excess rainfall from surface layer (Q1) in the upland field area was greatly increased, which resulted in lower discharge from subsurface layers (Q2 + Q3) and higher discharge from deeper layers (Q4 + Q5).

Conclusion

In this study, hydrological properties in an upland field catchment, which was reclaimed from a forest, were compared with a forest catchment. The geophysical conditions and soil properties of the surface and subsurface layers were investigated. The drainage system in the upland field catchment increases stream density and decreases the gradient of the field slope. The bulk density and water-holding capacity of soil in the upland field catchment were lower than that in the forest catchment, which result in decreased infiltration or percolation rate and water storage in the soil layers. The impacts of land use change on changes in hydrological factors were summarized as follows:

1) Infiltration rate was decreased immediately in the upland field area by rainfall and more excess rainfall produced in the upland field area.

2) Peak discharge in the upland field catchment increased 2.0–5.7 times compared to the forest catchment.

3) Direct runoff rate is smaller in the forest catchment (from 20.0% to 46.1%) than in the upland field catchment (from 29.8% to 57.1%). Additionally, it was found that the appearance of surface flow in the upland field catchment caused numerous rills and gullies, and increased soil erosion.

4) Annual evapotranspiration from the upland field showed a tendency to be smaller due to the change in vegetation type and available soil moisture for evapotranspiration.

Increase in peak discharge and surface runoff was considered as a negative impact on a forest catchment. It might further strengthen environmental stress through generating more sediment yield and erosion that were usually directly related to runoff volume and velocity. Therefore, the control of surface runoff producing can be considered as a primary hydrological component. A decline in infiltration rate and percolation rata would directly decrease recharge for the shallow and deeper aquifers. Such hydrological phenomena were considered a negative impact for the upland field catchment. This study will provide information for forest managers and decision makers in planning and making decisions for land and water resource management in the region.

References

[1] Crooks S, Davies H. Assessment of land use change in the Thames catchment and its effect on the flood regime of the river. Physics and Chemistry of the Earth, 2001, 26(7–8): 583–591

[2] Brath A, Montanari A, Moretti G. Assessing the effect on flood frequency of land use change via hydrological simulation (with uncertainty). Journal of Hydrology (Amsterdam), 2006, 324(1–4): 141–153

[3] Li Z, Liu W Z, Zhang X C, Zheng F L. Impacts of land use change and climate variability on hydrology in an agricultural catchment on the Loess Plateau of China. Journal of Hydrology (Amsterdam), 2009, 377(1–2): 35–42

[4] Wang Y, Takase K, He B. Application of a hydrologic model considering rainwater storage to analyze storm-induced landslides in a forest catchment. Water Resources Management, 2008, 22(2): 191–204

[5] Luo P, Takara K, Apip, He B, Nover D. Palaeoflood simulation of the Kamo River basin using a grid-cell distributed rainfall run-off model. Journal of Flood Risk Management, 2013 (in press)

[6] Costa M H, Botta A, Cardille J A. Effects of large-scale changes in land cover on the discharge of the Tocantins River, Southeastern Amazonia. Journal of Hydrology (Amsterdam), 2003, 283(1–4): 206–217

[7] Olchev A, Ibrom A, Priess J, Erasmi S, Leemhuis C, Twele A, Radler K, Kreilein H, Panferov O, Gravenhorst G. Effects of land-use changes on evapotranspiration of tropical rain forest margin area in Central Sulawesi (Indonesia): Modeling study with a regional SVAT model. Ecological Modelling, 2008, 212(1–2): 131–137

[8] Hernandez M, Miller S N, Goodrich D C, Goff B F, Kepner W G, Edmonds C M, Jones K G. Modeling runoff response to land cover and rainfall spatial variability in semi-arid watersheds. Environmental Monitoring and Assessment, 2000, 64(1): 285–298

[9] Ghaffari G, Keesstra S, Ghodousi J, Ahmadi H.SWAT-simulated hydroloigical impact of land-use change in the Zanjanrood Basin, Northwest Iran. Hydrological Processes, 2010, 24(7): 892–903

[10] Nie W M, Yuan Y P, Kepner W, Nash M S, Jackson M, Erickson C. Assessing impacts of landuse and landcover changes on hydrology for the upper San Pedro watershed. Journal of Hydrology (Amsterdam), 2011, 407(1–4): 105–114

[11] Lørup J K, Refsgaard J C, Mazvimavi D. Assessing the effect of land use change on catchment runoff by combined use of statistical tests and hydrological modeling: Case studies from Zimababwe. Journal of Hydrology (Amsterdam), 1998, 205: 147–163

[12] Niehoff D, Fritshch U, Bronstert A. Land-use impacts on storm-runoff generation: scenarios of land-use change and simulation of hydrological response in a meso-scale catchment in SW-Germany. Journal of Hydrology (Amsterdam), 2002, 267(1–2): 80–93

[13] Choi W, Deal B M. Assessing hydrological impact of potential land use change through hydrological and land use change modeling for the Kishwaukee River basin (USA). Journal of Environmental Management, 2008, 88(4): 1119–1130

[14] Takase K, Sato K. Comparison of evapotranspiration between a reclaimed upland field and a forest catchment. Journal of Japan Society of Hydrology and Water Resources, 1994, 7(6): 495–502

[15] Hong L, Takase K, Sato K. Properties of peak discharge and retention in a terraced paddy field catchment. Transactions of the Japanese Society of Irrigation, Drainage and Reclamation Engineering, 1998, 196: 189–195

[16] Wang Y, He B, Takase K. Effects of temporal resolution on hydrological model parameters and its impact on prediction of river discharge. Hydrological Sciences Journal, 2009, 54(5): 886–898

[17] Wang Y, He B, Takase K. Reply to the discussion of “Effects of temporal resolution on hydrological model parameters and its impact on prediction of river discharge” by Littlewood et al. Hydrological Sciences Journal, 2011, 56(3): 525–528

[18] Chow, V T. Hydrological determination of waterway areas for the design of drainage structures in small drainage basins. University of Illinois, Engineering Experiment Station, Bulletin 462, 1962

[19] Kadoya M, Fukushima A. Handbook of Rural Engineering. Japan Soc Irrigation, Drainage and Reclamation Engineering, 1989, 164 (in Japanese)

[20] Takeshita S. Takase K, Ihara K. Development of long-term runoff model including infiltration sub-model. Transactions of the Japanese Society of Irrigation. Drainage and Reclamation Engineering, 2001, 212: 71–77

[21] Nash J E, Sutcliffe J V. River flow forecasting through conceptual models part I—A discussion of principles. Journal of Hydrology (Amsterdam), 1970, 10(3): 282–290

[22] Gupta H J, Sorooshian S, Yapo P O. Status of automatic calibration for hydrological models: Comparison with multilevel expert calibration. Journal of Hydrologic Engineering, 1999, 4(2): 135–143

[23] Duan Q, Sorooshian S, Gupta V. Effective and efficient global optimization for conceptual rainfall-runoff models. Water Resources Research, 1992, 28(4): 1015–1031

[24] Duan Q, Sorooshian S, Gupta V. Optimal use of the SCE-UA global optimization method for calibrating watershed models. Journal of Hydrology (Amsterdam), 1994, 158(3–4): 265–284

[25] He B, Takase K, Wang Y. Regional groundwater prediction model using automatic parameter calibration SCE method for a coastal plain of Seto Inland Sea. Water Resources Management, 2007, 21(6): 947–959

[26] Wijesekara G N, Gupta A, Valeo C, Hasbani J G, Qiao Y, Delaney P, Marceau D J. Assessing the impact of future land-use changes on hydrological processes in the Elbow River watershed in southern Alberta, Canada. Journal of Hydrology (Amsterdam), 2012, 412–413: 220–232

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg