Assessing the hydrologic alteration of the Yangtze River using the histogram matching approach

Hydrologic changes of the Yangtze River, an important river with abundant water resources in China, were investigated using the Histogram Matching Approach. Daily streamflow data spanning the time interval from 1955 to 2013 was collected from Yichang and Datong stations, which monitor the hydrologic processes of the upper and lower reach of the Yangtze River, respectively. The Gezhouba Dam, the first dam constructed at the main stream of the Yangtze River, started operations in 1981. 1981 was used to differentiate the pre-dam (1955-1980) and post-dam (1981-2013) hydrologic regimes. The hydrologic regime was quantified by the Indicators of Hydrologic Alteration. The overall alteration degree of the upper Yangtze River was 31% and the alteration degree of every hydrologic indicator ranged from 10% to 81%. Only 1, 5 and 26 hydrologic indicators were altered at high, moderate and low degrees, respectively. The overall alteration degree of the lower Yangtze River was 30%, and the alteration degree of every hydrologic indicator ranged from 8% to 49%. No high alteration degree was detected at the Datong station. Ten hydrologic indicators were altered at moderate degrees and 22 hydrologic indicators were altered at low degrees. Significant increases could be observed for the low-flow relevant indicators, including the monthly flow from January-March, the annual minimum 1, 3, 7, 30 and 90-day flows, and the base flow index.


Introduction
The natural hydrological regime is a basic driver of water ecosystems and vitally important for keeping the health of river, lake and wetland ecosystems. It is affected by both anthropogenic activities and the climate, both of which have been the subject of many studies investigating worldwide hydrologic changes [1][2][3][4][5].
The Yangtze River plays an irreplaceable role in the ecological environmental protection and socio-economic construction of China, with its abundant water, hydropower, and biological resources. Over 50,000 dams/reservoirs have been built in the basin [6]. The total storage capacity of these dams is almost 180 billion m 3 , which amounts to approximately 20% of the annual runoff of the Yangtze River [7]. The impacts of these reservoirs on the hydrologic regime and its related ecological environmental implications have attracted considerable attention in recent years [8][9][10][11][12]. The present study seeks to evaluate the hydrologic changes of the Yangtze River using the Histogram Matching Approach (HMA). The results of this research will provide reference for water resource utilization and for the protection of riverine ecosystems.

Study site and data
The Yangtze River originates in the Qinghai-Tibet Plateau in western China (figure 1). Its length is about 6300 km and its catchment area is approximately 1.8 million km 2 . The basin lies in the monsoon region of East Asia, which is a subtropical climate zone. The monsoon-driven precipitation is unevenly distributed throughout the year, which causes seasonal variability in river runoff. The rainy season starts in April and determines the hydrologic year. The Yangtze River's hydrologic year begins from April and goes until March of the next calendar year. may not be wholly unaltered because some dams were constructed at the tributaries of the upper Yangtze River prior to 1981 [12,13]. This assessment aims to reveal the degree of alteration of the hydrologic regime from 1981 to 2013 compared with the hydrologic regime during the pre-dam era.

Indicators of Hydrologic Alteration (IHA)
The IHA method is widely applied to characterize the various components of hydrologic regimes [14]. It contains 33 hydrologic parameters that are grouped into five categories of hydrologic features. These are, group 1: magnitude of monthly water conditions, group 2: magnitude and duration of annual extreme water conditions, group 3: timing of annual extreme water conditions, group 4: frequency and duration of high and low pulses, and group 5: rate and frequency of water condition changes. The parameter "number of zero-flow days" was not applied because no zero-flow events occurred during the study period. Each IHA parameter was calculated for every hydrologic year, and the results were used to assess hydrologic alteration.

Histogram Matching Approach
The Range of Variability Approach (RVA) has been widely applied in worldwide assessments of hydrologic alteration and environmental flow design [15][16][17][18]. Shiau and Wu [19] have noted that the RVA method has two main limitations: (1) it only considers the frequency of the hydrologic parameters that fall within the target range, and ignores variation; and (2) it ignores the frequency and variation of the hydrologic parameters that fall above the upper-target or below the lower-target. These limitations might potentially lead to a false assessment of hydrologic alteration. The HMA method has been proposed to address the limitations of the RVA method. Its detailed procedure can be found in the literature of Shiau and Wu [19].
The hydrologic alterations were assessed by applying the HMA method. Then, following the class evaluation system [20], the degrees of hydrologic alteration for individual hydrologic indicators were classified into the categories "little alteration" (ranging from 0 to 33%, denoted by the letter "L"), "moderate alteration" (ranging from 34 to 67%, indicated by the letter "M"), and "high alteration" (ranging from 68 to 100%, indicated by the letter "H").

Results and discussion
The pre-dam and post-dam hydrologic regimes of the Yichang and Datong stations were compared in the histograms in figure 2 and figure 3, respectively, and the results of the hydrological alteration for every hydrologic indicator are shown in figure 4.
At the Yichang station, the overall degree of hydrologic alteration was 31%, which is classified as "little alteration." The deviations of the median flow in August, January, and March ranged from 37% to 43%, and this range is within the classification of "moderate alteration." The median flow deviations of the other months ranged from 10% to 31%, categorized as "little alteration." Figure 2 shows the change characteristics of the post-dam histograms as compared to the pre-dam histograms. The histograms of the median flow of January and March displayed similar change patterns. The postdam histogram moved to the right of the pre-dam histogram, indicating an increased flow in January and March. The alteration degree of magnitude and duration of annual extreme flow ranged from 10% to 33%, classified as "little alteration." Similar change patterns were observed in the annual minimum flow and the base flow index. The relative frequencies of the post-dam histogram were distributed to the right of the relative frequencies of the pre-dam histogram, indicating an increased flow in the dry season. The alteration degree of the Julian calendar date of the annual 1-day minimum flow was 35% ("moderate alteration") and the alteration degree of the Julian calendar date of the annual 1-day maximum flow was 13% ("little alteration"). The annual 1-day minimum flow usually occurred in January, February or March in the pre-dam years, but after 1981, it sometimes occurred in December of the previous calendar year ( figure 2 (24)). The alteration degree of frequency and the duration of high and low pulses fell about 20%, which is classified as "little alteration." The rise rate was found to have "moderate alteration" and the fall rate had "little alteration." The number of reversals showed "high alteration," and this was the only hydrologic indicator that exceeded 67% deviation. The number of reversals increased drastically after 1981.
At the Datong station, the overall degree of hydrologic alteration was 30%, which was very close to the alteration degree at the Yichang station. According to the data, no hydrologic indicators experienced "high alteration" at the Datong station. The median flows in January, February and March showed "moderate alteration", and during the other months the median flows showed "little alteration." The alteration degree of the annual minimum flow's relevant indicators ranged from 36% to 49% ("moderate alteration"). The annual maximum flow's relevant indicators ranged from 19% to 32% ("little alteration"). An increased flow in the dry season was also observed at the Datong station ( figure 3), which was similar to the change features at the Yichang station. All of the hydrologic indicators in Group 3 and 4 had "little alteration." The rise and fall rates showed "little alteration," but the number of reversals was found to have "moderate alteration." Different features of hydrologic alterations at the lower reach, as compared with hydrologic alterations at the upper reach, might have been influenced by the runoff from tributaries such as the Hanjiang River, as well as hydraulic regulation of the river channels and lakes between the Yichang and Datong stations, e.g. the Dongting and Poyang Lakes. The hydrologic regime of the Yangtze River represents a catchment response to climate input and anthropogenic activities. Precipitation may be a dominant factor in hydrologic changes. The spatial distribution of the precipitation regime can lead to spatial patterns of alterations in runoff [21]. Reservoir regulation is another important factor that affects the hydrologic regime. The significant increase of the low-flow relevant indicators may be caused by the water supply function of the reservoirs in the dry season [22][23][24]. The assessment results of the HMA method showed that the alteration degree of the hydrologic regime was affected by both climate change and anthropogenic activities, but this model was unable to definitively differentiate the individual roles of the two factors.  Figure 4. Hydrologic alterations of the Yangtze River evaluated using the HMA method.

Conclusions
The hydrologic alteration degree of the Yangtze River was assessed by applying the HMA method to long-term daily streamflow data from the Yichang and Datong stations. The overall alteration degree was approximately 30%, which is in the range of "little alteration." The alteration degree of the hydrologic indicators of the Yichang station ranged from 10% to 81%. 1, 5 and 26 hydrologic indicators were altered with high, moderate and low degree, respectively. The hydrologic indicator with the highest level of alteration was the number of reversals, which increased drastically after 1981. The alteration degree of the hydrologic indicators from the Datong station ranged from 8% to 49%. No hydrologic indicators showed high alteration. Ten hydrologic indicators exhibited "moderate alteration" and 22 hydrologic indicators showed "little alteration." Significant increases were observed for the low-flow relevant indicators, such as the monthly flow from January-March, the annual minimum 1-, 3-, 7-, 30-and 90-day flow, and the base flow index. The increased flow in the dry season in the post-dam period has helped to preserve the environmental and ecological water requirements, and also to protect the riverine ecosystem.