Climate change has been recognized as one of the major threats to the earth environment in the 21st century [1
]. Changes in temperature and precipitation patterns are expected to alter regional climates and hydrological systems [3
], affecting freshwater availability at the regional scale [4
]. The conflict of freshwater consumption between the human society and ecosystem may threat socio-economic sustainability and ecosystem health (e.g., leading to ecosystem degradation). Thus, improved understanding of water availability for both humans and ecosystems in the context of climate change is critical for better water resources management.
As far as the hydrological cycle is concerned, the composition of water resources includes blue water (BW) and green water (GW). Blue water is critical for domestic and industrial water consumption, while green water is crucial for supporting plant growth in rain-feed regions [5
]. With the development and improvement of hydrologic modeling, the spatial and temporal variation of surface and sub-surface hydrological components can be explicitly assessed by portioning water into Blue Water Flow (BWF; total water yield and deep aquifer recharge), green water flow (GWF; actual evapotranspiration) and green water storage (GWS; soil water content) [6
]. Long-term spatio-temporal changes in BW-GW can indicate the change in overall hydrological change at watershed scale.
Unlike the humid area, the semi-arid regions are more sensitive in streamflow reduction to climate change or anthropogenic water management, because of their remarkable seasonal climate characteristics. Climate warming in the semi-arid regions might cause increasing open water evaporation and plant transpiration [8
] which will further affect the formation of precipitation. Due to the limitation of surface water availability, understanding how the surface and ground water can be well managed in this region to meet the local urban agricultural and environmental demands under climate change is critical to water supply [9
]. Besides, seasonal differences, which are very important for agricultural and domestic water supply, are often neglected due to the concentration of precipitation in summer in these regions. Luanhe River Basin is a typical semi-arid region with uneven distribution of the precipitation. A total of 70%–80% of the annual precipitation falls in the rainy months of June to September [10
]. In addition, the southeast region received more precipitation than the northwest region with mean precipitation that varied from 400 to 800 mm from the northwestern to southeastern region. Therefore, the availability and temporal variability of hydrological components is of primary importance for fresh water planning and ecological security conservation in this basin.
Hydrological modelling has been employed to assess water availabilities in semi-arid and arid regions. Guo and Shen [11
] estimated water availability and agricultural water demand and proposed effective adaptation strategies to cope with severe water shortages under possible climate change trends in the arid region of northwestern China. Hydrological simulation and evaluation have also been conducted in Luanhe River Basin to evaluate the runoff change [10
], water scarcity [13
], and hydrological droughts under the climate change and human activities. Previous studies mainly focused on the integrated influences of LULC (Land Use/Land Cover) change and climate change, in which the individual effects of temperature and precipitation variations are often neglected. However, in Luanhe River Basin and other basins where vegetation is well protected, LULC has less influence on regional hydrological process. Hence, improved understanding of BW and GW under the influence of temperature and precipitation variability is of critical importance in the Luanhe River.
The present study was conducted in Luanhe River Basin, a semi-arid catchment in China. The objectives were as follows: (1) analyze the temporal variation of temperature and precipitation in the last five decades; (2) assess hydrological response to climate change in terms of different hydrologic components (BW and GW); (3) investigate the contribution of temperature and precipitation to the hydrological process.
5. Discussion and Conclusions
The impact of climate change on the acceleration of the hydrological cycle has gradually attracted increased attention. Based on the result of the SWAT model, this study discussed the variation trend of hydrological components under the impact of climate change in Luanhe River Basin during 1960–2017.
At the annual scale, the results of this study showed no significant trend in precipitation and hydrological components (BWF, GWF, and GWS) which was consistent with that of Zhang’s study [10
]. However, the novelty of this study resides in that we not only consider the influence of average temperature on hydrological process, but also take into account the influence of maximum and minimum temperature and their differences (DTR), which has been suggested to have a more significant effect on evapotranspiration [31
]. It is not sufficient to conclude that the influence of climate change on hydrological components were relatively small in the past 60 years, because the effects of climate elements on hydrological components may offset each other and had not emerged yet. Therefore, as suggested by Luo [18
], in the case of lacking a significant linear trend, decadal analysis was necessary and effective in the study of variation of hydrological components under the impacts of climate change. The increase of precipitation would lead to the increase of all three hydrological components, while the increase of DTR would reduce the increasing of GWF and decreasing of BWF and GWS.
At a seasonal scale, although there was no significant trend in the variation of precipitation and hydrological components in different seasons, there were distinct differences in their inter-generational variations. Moreover, with significant increasing of the temperature in all four seasons, we noticed that temperature, especially the lowest temperature (Tmin
) in winter, increased faster than that of other seasons, which resulted in a significant decrease in winter DTR. These trends were similar to results reported by Kowalczyk [33
], which also showed that the warming during the winter period was more evident than that in summer at the same study area. This study confirmed that unlike interannual variations, GWF and GWS might have the opposite trend with precipitation in spring, and the same trend in autumn and winter. This implied that the available water (rainfall) did not fully satisfy the atmospheric water demand in spring. Although influence trend of temperature was in accordance with the annual variation, the contribution rate of temperature changes to hydrological components varied significantly. Research had shown that the influence of climate change was especially concentrated in spring and winter, in which the influence was very complex [34
], such as the increasing of spring temperature would lead to the earlier beginning dates of growing season, and warming in winter would cause earlier snowmelt or precipitation falling as rain rather than snow, as mentioned in Vano’s research [35
]. Under the effected of these factors, uncertainty existed in the results, but it confirmed that both precipitation and temperature did have an important impact on hydrological processes in this region, especially in spring and winter [36
Climate change could be divided into four types: (I) DTR increase, humidity increase (2010s) (II) DTR decrease, humidity decrease (1980s); (III) DTR increase, humidity decrease (2000s); (IV) DTR decrease, humidity increase (1970s, 1990s). Manifestations about hydrological components of the different types of climate change were related to the response relationship. In one case, the effects of precipitation and temperature changes on hydrological components might be consistent with each other which was same as Lemann‘s opinions [37
]. Such as when the climate change type (I) occurred, the GWF would increase, whereas in type (II), the decrease of precipitation and DTR commonly promoted the decrease of GWF. When climate change type (III) occurred, BWF and GWS would decrease, while they would increase in type (IV). In another case, the effects of precipitation and temperature changes countered each other such as the changes of BWF and GWS in type (I) and (II) and that of GWF in type (III) and (IV), in which the change trend of hydrological components depended on the climate elements that have the larger contribution rate. Mostly precipitation dominated the change of BWF and GWS, while change of temperature had more contribution to the change of GWF, and with the increase of the temperature, its influence increased further. If the current climate trend continued in the future, as predicted by Hurst index, precipitation and DTR would decrease and temperature would rise, similar to 1980s, the circulating water in nature, especially GWF, would continue to decrease, and pose a great threat to plant growth [38
The study contributed to our understanding of the impacts of climatic factors on green and blue water variability in the Luanhe River Basin. As many studies [10
] have confirmed, the SWAT model could better interpret the changes of hydrological components in Luanhe River Basin, and the simulation accuracy were in good agreement with previous studies [13
]. However, there are certain uncertainties in the contribution analysis method. The meteorological series were already influenced by climate variability and land use change. Therefore, reorganization of meteorological data may change the relationship between meteorological elements, and further affect the calculation of contribution rates. There are several factors (e.g., short record, lack of hydrological stations) that may induce uncertainties and affect the simulation results. For example, due to the short length of calibration period, the simulation of the water balance component in climatically different periods may be subject to large uncertainties [41
]. In addition, the negligence of the building of hydraulic structures (reservoir, dam, and dike project) and the impact of groundwater exploitation also might have an influence on model accuracy and the water cycle. Therefore, future studies will focus more on the impacts of climate change and human activities including the construction of hydraulic engineering, irrigation, and exploiting groundwater on hydrological processes. Despite the uncertainty, the simulation accuracy of this model was satisfactory and can be used to evaluate the hydrological process in the study area. Results from this study are expected to help policymakers to better manage the water resources in the context of global and regional climate change. Relative impacts of precipitation and temperature in changing the dynamics of hydrological components, as quantified in this study, would help to adjust the urbanization, industrialization, and agriculture activities with the variability, and also could help to design planned climate change adaptation strategies for the governing factor in a more efficient and targeted way in this region.