•Identify and characterize drought events with a 3-dimensional view.•The spatio-temporal variation of drought in China during 1961–2012.•Merits and limitations of SPI, RDI, and SPEI.
Understanding ...the spatial and temporal variation of drought is essentially important in drought assessment. In most previous studies, drought event is usually identified in space and time separately, ignoring the nature of the dynamic processes. In order to better understand how drought changes have taken place in China during the past half-century, we carried out a comprehensive analysis of their spatio-temporal variation based on multiple drought indices from a climatic perspective. A 3-dimensional clustering method is developed to identify drought events in China from 1961 to 2012 based on the 0.25° gridded indices of SPI3 (3months Standardized Precipitation Index), RDI3 (3months Reconnaissance Drought Index) and SPEI3 (3months Standardized Precipitation Evapotranspiration Index). Drought events are further characterized by five parameters: duration, affected area, severity, intensity, and centroid. Remotely sensed soil moisture data were used to validate the rationality of identified drought events. The results show that the two most severe drought events in the past half century which occurred in the periods 1962–1963 and 2010–2011 swept more than half of the non-arid regions in China. Large magnitude droughts were usually centered in the region from North China Plain to the downstream of Yangtze River. The western part of North China Plain, Loess Plateau, Sichuan Basin and Yunnan-Guizhou Plateau had a significant drying trend, which is mainly caused by the significant decrease of precipitation. The three drought indices have almost the same performance in the humid regions, while SPI and RDI were found to be more appropriate than SPEI in the arid regions.
•The elasticity of runoff was derived based on the Choudhury–Yang equation.•Quantify impacts of changes in climate and land use/cover on the catchment runoff.•Attributing the dominant cause for ...runoff decrease in 33 hilly catchments in the Haihe River.
Catchment hydrological processes have been greatly influenced by the intensive variability in land use/cover, precipitation and air temperature due to climate change and local human activities. It is desired to understand catchment hydrological response to these changes. Observations show that annual runoff had a significant decreasing trend during the past 50years (1956–2005) in Haihe basin of northern China. In order to detect the major cause for this runoff decline, we first theoretically derived the elasticity of runoff from the Choudhury–Yang equation that is a water-energy balance equation based on the Budyko hypothesis. The elasticity of runoff was calculated in 33 selected mountainous catchments in Haihe basin based on their climate condition (represented by the aridity index, E0/P) and landscape condition (represented by the parameter, n). We analyzed the breakpoint of the annual runoff of the 33 catchments over the past 50years and split the whole study period into two sub-periods at the breakpoint (period 1: before the breakpoint; period 2: after the breakpoint). Then we attributed the runoff change between the two sub-periods to the impacts of climate variability and land use/cover change. The change of climate is represented by changes in precipitation (P) and potential evaporation (E0) and the change of land use/cover is represented by the parameter n in Choudhury–Yang equation. The change of annual runoff from period-1 to period-2 was the catchment hydrological response to the change of precipitation, potential evaporation and land use/cover (represented as ΔP, ΔE0 and Δn), and we calculated the runoff change based on the elasticities of runoff. For the 33 catchments, the mean annual runoff decreased by 43.0mm from the period-1 (91.4mm) to period-2 (48.4mm). Impacts of climate variation and land use/cover change were accountable for the runoff decrease by 26.9% and 73.1% on average, respectively. Impact of climate variation mainly came from the decrease in precipitation, and impact of land use/cover change mainly came from the vegetation increase. Vegetation increase was mainly due to the reforestation during the soil–water conservation practice during the past 30years and also partially due to climate variability especially the temperature increase. This methodology can also be used to predict the runoff change in these catchments without direct influence of local human activities under the future climate scenario based on the climate elasticity of runoff estimated from the historical hydroclimatic data.
•RNNs successfully simulated reservoir operations with multiannual flow regulation.•The applicability of the RNN-based operation under floods and droughts was tested.•A real-time operation combining ...RNNs and a hydrological model was developed.
Large-scale reservoirs play an essential role in water resources management for agriculture irrigation, water supply and flood controls. However, we need robust reservoir operation systems under both normal flow and extreme flow conditions. In this study, we applied recurrent neural networks (RNN) to simulate the operation of three multi-purpose reservoirs located in the upper Chao Phraya River basin. Two reservoirs have the function of multiannual flow regulation and one has the function of incomplete annual regulation. The goal of this study is to explore the applicability of RNN models for operation of reservoirs with multiannual flow regulation under different flow regimes, especially under extreme floods and droughts. We used three RNNs, namely nonlinear autoregressive models with exogenous input (NARX), long short-term memory (LSTM) and genetic algorithm based NAXR (GA-NAXR) for reservoir operation based on historical data. For real-time water resources management, an accurate inflow forecast is required to provide a real-time reservoir outflow, and thus we also carried out a real-time reservoir operation using the RNN and the inflow forecast by a distributed hydrological model. Results show that (1) GA-NARX has the highest accuracy among three RNNs and is more stable than the original NARX by optimizing the initial conditions, although it takes longer training time than NARX and LSTM; (2) GA-NARX-based operation model is effective under extreme floods and droughts; and (3) the real-time operation system combining the GA-NARX and the distributed hydrological model has reasonable accuracy in both wet season and dry season. RNN-based operation model developed in this study has potential applicability in practical water management, and the model combining the hydrological prediction is specially useful for real-time reservoir operation.
•We evaluated four latest satellite-precipitation products over the Yangtze River.•Multi-temporal evaluation focused on the context of hydrological applications.•We found apparent errors depend on ...regions, seasons and precipitation regimes.•Each satellite product has its own pros and cons.
In the present study, four high-resolution multi-sensor blended precipitation products, TRMM Multisatellite Precipitation Analysis (TMPA) research product (3B42 V7) and near real-time product (3B42 RT), Climate Prediction Center MORPHing technique (CMORPH) and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks (PERSIANN), are evaluated over the Yangtze River basin from April 2008 to March 2012 using the gauge data. This regional evaluation is performed at temporal scales ranging from annual to daily, based on a number of diagnostic statistics. Gauge adjustment greatly reduces the bias in 3B42 V7, a post real-time research product. Additionally, it helps the product maintain a stable skill level in winter. When additional indicators such as spatial correlation, Root Mean Square Error (RMSE), and Probability of Detection (POD) are considered, 3B42 V7 is not always superior to other products (especially CMORPH) at the daily scale. Among the near real-time datasets, 3B42 RT overestimates annual rainfall over the basin; CMORPH and PERSIANN underestimate it. In particular, the upper Yangtze always suffers from positive bias (>1mmday−1) in the 3B42 RT dataset and negative bias (−0.2 to −1mmday−1) in the CMORPH dataset. When seasonal scales are considered, CMORPH exhibits negative bias, mainly introduced during cold periods. The correlation between CMORPH and gauge data is the highest. On the contrary, the correlation between 3B42 RT and gauge data is more scattered; statistically, this results in lower bias. Finally, investigation of the probability distribution functions (PDFs) suggests that 3B42 V7 and 3B42 RT are consistently better at retrieving the PDFs in high-intensity events. Overall, this study provides useful information about the error characteristics associated with the four mainstream satellite precipitation products and their implications regarding hydrological applications over the Yangtze River basin.
Climate elasticity of runoff is an important indicator for evaluating the effects of climate change on runoff. Consequently, this paper proposes an analytical derivation of climate elasticity. Based ...on the mean annual water‐energy balance equation, two dimensionless numbers (the elasticities of runoff to precipitation and potential evaporation) were derived. Combining the first‐order differential of the Penman equation, the elasticities of runoff to precipitation, net radiation, air temperature, wind speed, and relative humidity were derived to separate the contributions of different climatic variables. The case study was carried out in the Futuo River catchment in the Hai River basin, as well as in 89 catchments of the Hai River and the Yellow River basins of China. Based on the mean annual of climatic variables, the climate elasticity in the Futuo River basin was estimated as follows: precipitation elasticity , net radiation elasticity , air temperature elasticity , wind speed elasticity , and relative humidity elasticity . In this catchment, precipitation decrease was mainly responsible for runoff decline, and wind speed decline had the second greatest effect on runoff. In the 89 catchments of the Hai River and the Yellow River basins of China, climate elasticity was estimated as follows: ranging from 1.6 to 3.9, ranging from −1.9 to −0.3, ranging from −0.11 to −0.02°C−1, ranging from −0.8 to −0.1, and ranging from 0.2 to 1.9. Additional analysis shows that climate elasticity was sensitive to catchment characteristics.
Key Points
Proposes an analytical derivation of climate elasticity
Elasticity of runoff to precipiation, net radiation, temperature, etc.
Climate elasticity of runoff over Northern China
•Identify and characterize drought events spatio-temporally.•Copula based joint probability distribution of duration, area, and severity.•Drought return period estimation in Southwest China.•Evaluate ...the sensitivity of the 3-D copulas based drought frequency analysis.
Drought frequency analysis is a prerequisite for drought resistance planning and drought risk management. Drought is a spatio-temporal dynamic process, usually characterized by its duration, spatial extent, and severity. Copula based multivariate frequency analysis has been widely used to calculate drought frequency. However, the spatial extent is scarcely considered in previous studies, due to the fact that drought event is usually identified either for a fixed spatial scale or for a fixed temporal scale. This study develops a regional drought frequency analysis model based on trivariate copulas by considering the spatio-temporal variations of drought events. Drought duration, drought affect area, and drought severity are identified first, and their trivariate joint distribution is constructed later. The model is applied for drought frequency analysis in Southwest China during 1961–2012. A variety of probability distribution functions and copula functions (including elliptical, symmetric and asymmetric Archimedean) are used as candidate choices, and the most appropriate ones are selected based on goodness of fit using different methods. The robustness of drought frequency analysis is then evaluated and discussed. The results show that drought frequency analysis needs to fully consider the three characteristic parameters (duration, affect area, and severity) reflecting drought spatio-temporal variability. And the drought return period estimated by the copula-based trivariate frequency analysis appropriately integrates the effects of drought duration, affect area and severity, which is a reliable drought statistical measurement. The 2009–2010 drought, which has a return period of about 94years, is the most severe one in Southwest China during the period of 1961–2012. The Joe and Gumbel copulas are found to be more suitable to estimate the joint distribution of drought duration, affect area and severity, and the Asymmetric (nested) function forms perform better than the symmetric functions.
•This water cycle study focuses on the impacts of climate change and human activities.•River basin management requires an integrated model of hydro-bio-geochemistry.•Co-evolution of the human–water ...systems should be the focus of future study.
Water is the fundamental natural resource that supports life, ecosystems and human society. Thus studying the water cycle is important for sustainable development. In the context of global climate change, a better understanding of the water cycle is needed. This study summarises current research and highlights future directions of water science from four perspectives: (i) the water cycle; (ii) hydrologic processes; (iii) coupled natural-social water systems; and (iv) integrated watershed management. Emphasis should be placed on understanding the joint impacts of climate change and human activities on hydrological processes and water resources across temporal and spatial scales. Understanding the interactions between land and atmosphere are keys to addressing this issue. Furthermore systematic approaches should be developed for large basin studies. Areas for focused research include: variations of cryosphere hydrological processes in upper alpine zones; and human activities on the water cycle and relevant biogeochemical processes in middle-lower reaches. Because the water cycle is naturally coupled with social characteristics across multiple scales, multi-process and multi-scale models are needed. Hydrological studies should use this new paradigm as part of water-food-energy frontier research. This will help to promote interdisciplinary study across natural and social sciences in accordance with the United Nation's sustainable development goals.
Display omitted
►
E
pan over China showed a −3.1
mm/a
2 (−1.8% per decade) change from 1961 to 2001. ► We proposed the PenPan-20 model for evaporation from the Chinese micro-pan. ► We quantified the attribution of ...changing
E
pan based on the PenPan-20 model. ► Main causes of changing
E
pan were declining wind speed in most parts of China. ► Main causes of changing
E
pan were reducing solar radiation in southwest China.
Pan evaporation (
E
pan) is an indicator of atmospheric evaporative demand, and downward trends have been reported in many regions over the past several decades. It is important to understand why
E
pan has changed and determine what the main causes are. This study analyzed daily climate data from 54 meteorological stations across China measured from 1961 to 2001. Climatic factors included
E
pan measured with the standard Chinese 20
cm diameter pan, global solar radiation (
R
s), air temperature (
T
a), relative humidity (RH), and wind speed (
U). We modified the PenPan model for pan evaporation from the Class A pan, to estimate Chinese micro-pan evaporation
E
pan and quantify the contributions of climatic factors (
R
s, net longwave radiation
R
nl,
T
a,
U, and vapor pressure deficit
D) to
E
pan using partial derivatives. For China as a whole, the arithmetic average of
E
pan from these 54 stations showed a significant decline of −3.1
mm
a
−2 (−1.8% per decade). The observed change in
R
s was −4.5
W
m
−2 decade
−1 (−2.6% per decade); the mean
T
a increased by 0.27
°C per decade;
U decreased by −0.10
m
s
−1 decade
−1 (−6% per decade); and
D increased by 0.012
kPa per decade (2% per decade). For China as a whole, the dominant factors influencing changing
E
pan were
R
s and
U; the causes of
E
pan were very varied: regionally, the main causes for change were in most parts of China, it was the change in
U, in southwest China it was the change in
R
s.
•A frequency-based method to determine flood threshold by results of distributed model.•Using binary classification to derive rainfall threshold for flood warning.•Relationship between rainfall ...threshold and soil moisture state varies with climate.•Applicability of the method to both gauged and ungauged catchments.
Flash flooding is one of the most common natural hazards in China, particularly in mountainous areas, and usually causes heavy damage and casualties. However, the forecasting of flash flooding in mountainous regions remains challenging because of the short response time and limited monitoring capacity. This paper aims to establish a strategy for flash flood warnings in mountainous ungauged catchments across humid, semi-humid and semi-arid regions of China. First, we implement a geomorphology-based hydrological model (GBHM) in four mountainous catchments with drainage areas that ranges from 493 to 1601km2. The results show that the GBHM can simulate flash floods appropriately in these four study catchments. We propose a method to determine the rainfall threshold for flood warning by using frequency analysis and binary classification based on long-term GBHM simulations that are forced by historical rainfall data to create a practically easy and straightforward approach for flash flood forecasting in ungauged mountainous catchments with drainage areas from tens to hundreds of square kilometers. The results show that the rainfall threshold value decreases significantly with increasing antecedent soil moisture in humid regions, while this value decreases slightly with increasing soil moisture in semi-humid and semi-arid regions. We also find that accumulative rainfall over a certain time span (or rainfall over a long time span) is an appropriate threshold for flash flood warnings in humid regions because the runoff is dominated by excess saturation. However, the rainfall intensity (or rainfall over a short time span) is more suitable in semi-humid and semi-arid regions because excess infiltration dominates the runoff in these regions. We conduct a comprehensive evaluation of the rainfall threshold and find that the proposed method produces reasonably accurate flash flood warnings in the study catchments. An evaluation of the performance at uncalibrated interior points in the four gauged catchments provides results that are indicative of the expected performance at ungauged locations. We also find that insufficient historical data lengths (13years with a 5-year flood return period in this study) may introduce uncertainty in the estimation of the flood/rainfall threshold because of the small number of flood events that are used in binary classification. A data sample that contains enough flood events (10 events suggested in the present study) that exceed the threshold value is necessary to obtain acceptable results from binary classification.
Many previous studies have evaluated the hydrologic response to climate change using the first‐order approximation (first‐order Taylor expansion) of the Mezentsev‐Choudhury‐Yang equation (formulating ...the Budyko hypothesis), which has a parameter n representing catchment characteristics. However, no studies have paid attention to the error due to the first‐order approximation. This study therefore estimates this error to improve the theoretical framework for assessing the contribution of climate change to runoff based on the Budyko hypothesis. Specifically, the error increases when precipitation (P) decreases and potential evaporation (E0) increases, and n increases. Therefore, an increasing P or decreasing E0 leads to an underestimate of the climatic contributions, while a decreasing P or increasing E0 leads to an overestimate. In addition, we suggest a new method to accurately estimate the contribution of climate change to runoff.
Key Points:
Error analysis of the first‐order Taylor expansion of Budyko hypothesis
A new method for assessing hydrologic response to climate change