Advances in Meteorology

Volume 2016 (2016), Article ID 1404290, 13 pages

http://dx.doi.org/10.1155/2016/1404290

## Impact of Climate Change on Hydrologic Extremes in the Upper Basin of the Yellow River Basin of China

^{1}College of Hydrology and Water Resources, Hohai University, Nanjing 210098, China^{2}State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering, Hohai University, Nanjing 210098, China

Received 16 October 2015; Revised 21 February 2016; Accepted 17 March 2016

Academic Editor: Jingfeng Wang

Copyright © 2016 Jun Wang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

#### Abstract

To reveal the revolution law of hydrologic extremes in the next 50 years and analyze the impact of climate change on hydrologic extremes, the following main works were carried on: firstly, the long duration (15 d, 30 d, and 60 d) rainfall extremes according to observed time-series and forecast time-series by dynamical climate model product (BCC-CSM-1.1) were deduced, respectively, on the basis that the quantitative estimation of the impact of climate change on rainfall extremes was conducted; secondly, the SWAT model was used to deduce design flood with the input of design rainfall for the next 50 years. On this basis, quantitative estimation of the impact of climate change on long duration flood volume extremes was conducted. It indicates that (1) the value of long duration rainfall extremes for given probabilities (1%, 2%, 5%, and 10%) of the Tangnaihai basin will rise with slight increasing rate from 1% to 6% in the next 50 years and (2) long duration flood volume extremes of given probabilities of the Tangnaihai basin will rise with slight increasing rate from 1% to 6% in the next 50 years. The conclusions may provide technical supports for basin level planning of flood control and hydropower production.

#### 1. Introduction

Climate change has affected many fields of nature and human society in recent years and has been one of the most attractive research fields. On the background, the estimation and simulation of the impact of climate change in the hydrology have been becoming a research topic. Many hydrologists [1–4], who have done research on the hydrological response to climate change and human activities, believe that anthropogenic global climate change and human activities have significantly affected hydrologic cycle and resulted in changes in the spatial and temporal distribution of water resources at both global and local scales. It is no doubt that the changes in the hydrological cycle will have serious impacts on ecological, social, and economic situations [5, 6] and bring us severe challenges. To respond to the climate change challenge, one of the most important tasks is to reveal influencing mechanism of hydrologic cycle by climate change and to predict the impact on corresponding fields. Many hydrologists have been working on it; for example, some studies have identified robust trends over some specific regions [7, 8]. However, there are still many questions unclear and the research should be continued. With the development of society, the demand of scientificity of the basin level and national planning of flood control, hydropower production, agricultural irrigation, and ecosystem preservation are increasing; hence, projecting the future climate and assessing its probable impact on water resources are critical. Many studies on the impacts of climate change on hydrological regimes [9–11] have been conducted. In these studies, global climate models and hydrologic model were usually used to simulate the changes in hydrological regimes at watershed scales.

Most of the previous climate change impact assessment studies on hydrological processes of Yellow River watershed focused on the trend of hydrologic elements [12]. However, the assessment study on revolution law of the hydrologic extremes responding to climate change of Yellow River basin is hardly a blank. To reveal the revolution law of hydrologic extremes and assess the impact of climate change on hydrologic extremes in the upper Yellow River basin, the objectives of the paper include (1) evaluation of the impact of climate change on precipitation extremes of long duration (15 d, 30 d, and 60 d) for given probabilities (1%, 2%, 5%, and 10%); (2) evaluation of the impact of climate change on flood volume extremes of long duration (15 d, 30 d, and 60 d) for given probabilities (1%, 2%, 5%, and 10%), through hydrologic frequency analysis that deduce the design rainfall for the measured phase and the future by running distributed hydrological model (SWAT) with the input condition of design rainfall to deduce the flood volume extreme of different duration. The paper focuses on revealing the impact of climate change on the hydrologic extremes of the Tangnaihai basin which locates in the upstream of Yellow River; the main content is as follows: the methods applied to study the impact of climate change on hydrologic extreme, including hydrologic frequency analysis method, the bias correction method, and the hydrological model, are described in Section 2. The study area and available data are then introduced in the Section 3. The results of impact of climate change on hydrologic extremes are presented in Section 4. The conclusions are finally remarked in Section 5.

#### 2. Methodologies

##### 2.1. Frequency Analysis of Precipitation Extremes

To reveal the statistical law of the hydrologic extremes, hydrologic time series analysis and modeling are an effective approach. Obviously, hydrologic frequency analysis is one of most popular approaches which are based on time series to analyze the law of hydrologic extremes [13, 14]. The key of hydrologic frequency analysis is to determine the probability distribution of extreme. As we known, there are many functions, including extreme value distribution (Gumbel distribution), generalized extreme value distribution (GEV), log-normal distribution (L-N), the Pearson type III distribution (P-III), and the logarithmic Pierre Johnson III distribution, that could be used as the probability distribution of hydrologic extremes.

In China, Liang [15] showed that the P-III distribution is suitable for description of the statistical law of hydrologic extremes, such as annual maximum rainfall, annual maximum flood peak discharge, and annual maximum flood volume of different duration, based on the application experiences. Therefore, the Pearson type III (P-III) curve has been used for the hydrologic frequency analysis in China.

The P-III curve is known as the distribution mathematically (Gamma distribution with three parameters). Its probability density function is expressed as follows:where , , and denote the position, scale, and shape parameters of the distribution, respectively. The relation of these parameters and three moments (, , and ) can be expressed as follows:where denotes the mean of hydrologic extreme time series; is the variance of hydrologic extreme time series; and is the variable coefficient of hydrologic extreme time series.

Hydrologic frequency analysis and calculation are to ascertain the random variable corresponding to the specified frequency , which can be obtained by the distribution function defined by transcendental probability:

To simplify the integration solution of (3), the variables of the Pearson type III distribution can be obtained by standard transformation of variable :where is known as coefficient of mean deviation. Then the integral operation of iswhere the integrand contains only one unknown parameter or (). According to hydrologic customary, the relationship of , , and is tabulated in advance, namely, -value hydrographic table. The corresponding can be obtained through the inverse transform of (4) which is expressed as follows:

##### 2.2. Frequency Analysis of Flood Volume Extremes with Different Durations

Approach to deduce flood volume extremes with different durations can be divided into two types according to the data condition [16]: one is the so-called direct method when the length of observed discharge time series is relatively long. According to observed maximum discharge time series, the flood volume extremes of different duration with a certain probability can be deduced by hydrologic frequency analysis as shown in Section 2.1; the other one is the so-called indirect method when the length of observed discharge time series is relatively short or there is no observed discharge data. According to observed precipitation data, the precipitation extremes of different duration with a certain probability can be deduced by hydrologic frequency analysis firstly, on the basis that the flood volume extremes with different durations of corresponding probability can be deduced by rainfall-runoff model on the hypothesis that rainfall with given frequency could generate the flood with the same frequency.

There are many hydrologic models for rainfall-runoff simulation; in the paper, the SWAT model which is famous as distributed hydrologic model and widely applied all over the world was adopted to deduce the flood volume.

##### 2.3. Hydrologic Model: SWAT

To evaluate quantitatively the impact of climate change on the flood extreme, hydrologic model is needed to simulate the flood. There are many rainfall-runoff models such as Xin’anjiang model, TOPMODEL, and VIC model. In the paper, SWAT model was selected because of its powerful hydrological process simulation capabilities. Known as a famous distributed hydrologic model, SWAT model is a continuous-time, semidistributed, process-based river basin or watershed scale model. SWAT model was developed to predict the impact of land management practices on water, sediment, and chemical yields in agricultural watersheds with varying soils, land use, and management conditions over long period of time [17, 18]. Comparing with other hydrologic models, SWAT has two outstanding features. One is the use of Hydrologic Response Unit (HRU), which is divided according to land use, soil distribution, and slope type, as the calculation unit [19]. SWAT divides a watershed into subbasins. Each subbasin is connected through a stream channel and further each subbasin is divided into HRUs. SWAT simulates hydrology and sediment at the HRU level. Water and sediment from each HRU are summarized in each subbasin and then routed through the stream network to the watershed outlet [18, 20]. The other one is the simulation of surface runoff by using the modified SCS curve number which is deduced based on land use and soil type of watershed [21].

##### 2.4. Bias Correction of Climate Model

Systematic errors of climate models may lead to unrealistic hydrological simulations of river flow [22, 23]; thus, bias correction methods must be implemented to correct the climate product before application and analysis. For adjusting climate model product, the linear scaling, local intensity scaling, power transformation, variance scaling, distribution transfer, and the delta-change approaches are the commonly used bias correction methods [24]. Bias correction methods are based on the assumption that the same correction algorithm applies to both current and future climate conditions. In the study, linear scaling method was employed to correct the daily precipitation and monthly precipitation of climate model.

Based on the precipitation simulated by climate model and the corresponding measured precipitation, correction coefficients of each month were calculated. The precipitation predicted by climate model was modified on the basis of correction coefficients scaling monthly precipitation data predicted by the model. The correction coefficient can be calculated bywhere is the correction coefficient of the th month, is the measured monthly mean precipitation of the th month of reference period, and is the monthly mean precipitation simulated by the climate model of the th month of reference period.

#### 3. Study Area and Base Data

##### 3.1. Study Area

Tangnaihai basin with a drainage area of 122,000 km^{2} is located in the upstream of Yellow River in the western China, accounting for 15% of that of the Yellow River basin. Annual average runoff amount at the Tangnaihai cross section is 205.2 × 10^{8} m^{3}, accounting for 40% of annual mean runoff amount of Yellow River basin. It is a semihumid region with good vegetation and less human activity. In the study, nine meteorological stations and seven hydrologic stations are involved. Figure 1 shows the location of the study area and gauging stations distribution.