Assessment of hydro-climatological drought conditions for Hong-Thai Binh river watershed in Vietnam using high-resolution model simulation

Understating hydro-climatological conditions in a transboundary is always challenging because of issues in sharing available data among riparian countries. The present study has explored the hydro-climatological drought conditions over Hong-Thai Binh river watershed (H-TBRW) based on the downscaled rainfall and reproduced streamflow by the state-of-the-art coupled regional hydroclimate model. The standardized precipitation index (SPI) and streamflow drought index (SDI) indicators are used to define the climatological and hydrological drought conditions, respectively. Both SPI and SDI are derived from the precipitation and streamflow data reproducibility for the H-TBRW during 1950-2015. The results demonstrate a slight increasing trend in both climatological and hydrological conditions. Over the H-TBRW, results reveal that the Da and Thao rivers strongly expect drought conditions; meanwhile, the remaining rivers are very likely to experience similar drought conditions as in the past.

Environmental Sciences | Climatology Doi: 10.31276/VJSTE.61(2).90-96 Assessment of hydro-climatological drought conditions for Hong-Thai Binh river watershed in Vietnam using high-resolution model simulation Ho Viet Cuong1, Do Hoai Nam1*, Trinh Quang Toan2 Vietnam Academy for Water Resources Hydrologic Research Laboratory, Department of Civil and Environmental Engineering, University of California, USA Received 20 March 2019; accepted June 2019 Abstract: Understating hydro-climatological conditions in a transboundary is always challenging because of issues in sharing available data among riparian countries The present study has explored the hydro-climatological drought conditions over Hong-Thai Binh river watershed (H-TBRW) based on the downscaled rainfall and reproduced streamflow by the state-of-the-art coupled regional hydroclimate model The standardized precipitation index (SPI) and streamflow drought index (SDI) indicators are used to define the climatological and hydrological drought conditions, respectively Both SPI and SDI are derived from the precipitation and streamflow data reproducibility for the H-TBRW during 1950-2015 The results demonstrate a slight increasing trend in both climatological and hydrological conditions Over the H-TBRW, results reveal that the Da and Thao rivers strongly expect drought conditions; meanwhile, the remaining rivers are very likely to experience similar drought conditions as in the past Keywords: coupled WEHY-HCM model, drought, Standardized Precipitation Index, Streamflow Drought Index Classification number: 5.2 Introduction In the monsoon regions, though annual mean rainfall is high, the rainfall distribution is quite distinct between the seasons The rainy season often accounts for 70-90% of the annual mean rainfall [1] Under a changing climate, increases in surface temperature tend to accelerate evapotranspiration processes, causing greater water vapour in the air that subsequently results in more precipitable water However, increased precipitation is mostly distributed in the wet season; meanwhile, the dry season is very likely to be drier (e.g., [2, 3]) In other words, droughts are intensifying and are causing adverse impacts on lives, water resources, agriculture, and food security Conventional assessments of trend and variability of droughts were mostly conducted using ground hydrometeorological observation (e.g., [4, 5]) or combined observation and model simulations [6] It is known that the existing ground observation networks in developing countries are quite scattered and are extremely short on record length This situation diminishes studies of drought conditions, especially the investigation of spatial variation of droughts across transboundary river basins where data are inaccessible or are not shared among the riparian countries As an extension of the previous work regarding the reconstruction and evaluation of changes in hydrologic conditions over a transboundary region [7], this study will further capture the trend and variability of droughts in the past climate (1950-2015) The work will be based on the simulations derived from a regional climate model coupled with a physically based hydrology model for the H-TBRW, the portion lying in the territory of Vietnam of the Red river Some well-known drought indices are employed to detect the trend and variability of both meteorological and hydrological drought conditions These indices are calculated for a range of time scales as addressed in the literature (e.g., [3, 5]) in order to provide a choice of index appropriate for different meteorological, agricultural and hydrological applications Methodology, study area, and data *Corresponding author: Email: namdh@vawr.org.vn 90 Vietnam Journal of Science, Technology and Engineering JUne 2019 • Vol.61 Number Hydro-meteorological drought indicators Droughts often cause impacts over a widespread area 11 22 2.00orormore more Extreme Extremewet wet 0toto-0.99 -0.99 2.00 State SPI value Positive Category 1.50toto1.99 1.99 Severe Severewet wet 1.50 Positive 2.00 or more Extreme wet Milddrought drought Mild SPI value Negative -1.00 to -1.49 -1.00 to -1.49 Negative to -0.99 Category Moderatedrought drought Moderate Mild drought 1.002.00 Moderate wet 0-1.50 -1.99 MildSevere Severedrought drought 3 12 1.00 toto1.49 wet toto-1.99 or1.49 more Moderate Extreme to-1.50 1.50 to 1.99 Severe wetwet -1.00 to-0.99 -1.49 Moderatedrought drought to 1.49 1.99 Mild Severe wet -1.00 Moderate drought 0.99 Mild wet -2.00 orless lessSevere Extreme drought 4 23 0toto1.50 0.99 wet -2.00 or-1.49 Extreme drought 1.00 to Moderate wet -1.50 toto-1.99 drought | 34 Environmental Sciences Climatology 1.00 to 1.49 Mild Moderate -1.50ortoless -1.99 Extreme Severedrought drought to 0.99 wet wet -2.00 Streamflow Drought Index (SDI): similar to SPI, theSDI SDIwas wasdeveloped developedtoto Streamflow Drought Index (SDI): similar to SPI, the 0Streamflow to 0.99 Mild wet Index -2.00 or less Extreme drought Drought (SDI): similar to SPI, the SDI was developed to explore the water resources conditions of the watershed based on the information explore the water resources conditions of the watershed based on the information ofof explore the water resources conditions of thesimilar watershed basedthe the information streamflow volumes for periods, asonexpressed expressed inof to [15]: during a long period of time; these are commonly referred to cumulative Drought Indexreference to SPI, SDI was developed streamflow volumes for(SDI): reference periods, equation cumulative Streamflow streamflow volumes for reference periods, asasexpressed ininequation [15]: cumulative volumes for reference periods, as expressed [15]: of 2water [15]: as social, economic, and social impacts It is widely accepted equation explore thestreamflow resources conditions of the watershed based in onequation the information that droughts are defined in terms of meteorological, cumulative streamflow volumes for reference periods, as expressed in equation(2) 2(2)[15]: (2) ∑∑ (2) ∑ hydrological, agricultural, and socioeconomic conditions ∑ the cumulative However, this study considers only the first two terms, and where denotes the cumulative streamflow for where Vi,k denotes streamflow volume for thefor i-th hydrological yearthe and (2) year where Vdenotes i,k thecumulative cumulative streamflow volume forvolume thei-th i-thhydrological hydrological yearand and Vi,ki,kdenotes the streamflow volume the where V drought indices are calculated solely based on precipitation i-th hydrological and the k-th reference period, k =k =13 the k-th reference period,year k = for October-December, k = for October–March, thek-th k-threference reference period, for October-December, 2for forOctober–March, October–March, the cumulative streamflow volume for thekki-th year and kk==33 where Vi,k denotes period, kk==11for ==2hydrological and streamflow data as described in the followingthe for October-December, k =October-December, for October-March, k = for for October–June, and k = for October-September the k-th reference period, k = for October-December, k = for October–March, k=3 for October–June, and k = for October-September paragraphs for October–June, andand k = k4 for October-June, = 4October-September for October-September Based on cumulative streamflow volumes Vi,k, the SDI is defined for each for October–June, and k = for October-September Standardized Precipitation Index (SPI): precipitation , the is for Based on cumulative streamflow volumes Vthe SDIisSDI isdefined defined foreach each Based onkcumulative cumulative streamflow volumes reference period of the i-th hydrological year as follows: Based on streamflow volumes VVi,ki,k, ,the i,k SDI SDIhydrological is defined for each on cumulative streamflow volumes Vthe i,k, the and evapotranspiration are primary variables controlling definedBased for each reference period k of i-th referenceperiod periodkkofofthe the i-thhydrological hydrologicalyear yearasasfollows: follows: reference period k of thei-th i-th hydrological year as follows: (3) the formation and persistence of drought conditions.reference year as follows: However, it is quite difficult to estimate evapotranspiration (3) (3) (3) where Vk and k are the mean and the standard deviation of cumulative streamflow (3) rates, so drought climatology studies have used mostly volumes of reference period k as these are estimated over a long period of time, data on precipitation Among the available indices in the where the mean and the standard deviation Vk and kdefinition, kare areσthe theare mean standard deviation of cumulative streamflow where Vand k and the meanand andthe the standard deviation ofcumulative cumulative where Vk kand respectively this SDI values are also categorized into five states of streamflow By mean and the standard deviation ofperiod literature used to identify meteorological drought conditionwhere k kare streamflow of Vcumulative volumes of reference k as streamflow volumes of reference period k as these are estimated over a long period of time, of time, hydrological conditions of the watershed as presented in Table - for example, Palmer drought severity index [8], cropvolumes volumes ofreference reference period theseperiod areestimated estimated over longperiod period theseofare estimated over aaslong of time,over respectively period kkas these are a along of time, respectively By this definition, SDI values are also categorized into five states of moisture index [9], and surface water supply index [10] - respectively By this definition, SDI values are also categorized into Bythis thisdefinition, definition,SDI SDIvalues valuesare arealso alsocategorized categorizedinto into fivestates statesofof respectively By hydrological conditions of the watershed as presented in Table 2.watershed asfive the standardized precipitation index (SPI) has been widely five states of hydrological conditions of the hydrologicalconditions conditions ofthe thewatershed watershedasaspresented presentedininTable Table2.2 accepted for drought assessment studies (e.g., [11-14]) Thehydrological presented in Tableof SPI is formulated to estimate the precipitation deficit for multiple time scales, i, which indicate drought conditions Table Drought classification by SDI value (modified after 33 [15]) throughout the watershed State SPI value Category SPIis issimply simply defined as the of difference the difference precipitation from the SPI defined as the ratioratio of the of of precipitation from the mean for a specified time period over mean for a specified time period over the corresponding standard deviation Positive determined the corresponding standard deviation determined from past Greater than Non-drought from past records as expressed in equation below: records as expressed in equation below: zero (1) (1) SPI value Category Negative to -0.99 Mild drought -1.00 to -1.49 Moderate drought -1.50 to -1.99 Severe drought where is standardizedprecipitation precipitation index index for -2.00 or less Extreme drought where SPISPI is standardized fortime timescale scale i (e.g., 1-, 3-, 6-, 12-, 24-, i (e.g., 1-, 3-, 6-, 12-, 24-, and 48-month time scales); Pi for time mean scale i; is climatological mean andis48-month timefor scales); Pi is precipitation Study area precipitation time scale i; Pi is climatological precipitation is standard deviation deviation precipitation timeriver scaleisi categorized among the five major precipitation for for timetime scalescale i; i;is sstandard TheforRed i precipitation for time scale i transboundary river systems in Southeast Asia and flows The SPI is computed by fitting a probability density function to the frequency The SPI is computed by fitting a probability density from Yunnan province in Southwest China through northern distribution summed over the scale of interest values can be (Fig 1) The Red river covers VietnamSPI to the Gulf of Tonkin function of to precipitation the frequency distribution of time precipitation a drainage area of 169,020 greater (positive) or time less (negative) than the precipitation Table 1km2, of which 48% is in China’s summed over the scale of interest SPIclimatological values can be mean territory, 51% in Vietnam’s territory, and only 1% is in greater (positive) or lessSPI (negative) the climatological below depicts categorical valuesthan reflecting drought classifications fromisextremely mean precipitation Table below depicts categorical SPI Laos’ territory The H-TBRW is named for the downstream wet to dry conditions values reflecting drought classifications from extremely wet portion of the Red river basin in Vietnam The H-TBRW covers to dry conditions.classification by SPI value (modified after Table Drought [5]) 26 provinces and cities (including Hanoi and Hai Phong), with a total population of 30 million Table Drought classification by SPI value (modified after [5]) State SPI value Category SPI value Category As it is located in a tropical region, the H-TBRW is strongly influenced by the tropical monsoon climate Positive Negative Positive Negative Average annual precipitation is spatially distributed in a 2.00 or more Extreme wet to -0.99 Mild drought wide range over the river basin (from 700-2,100 mm in 2.00 or more Extreme wet to -0.99 Mild drought China to 1,200-4,800 mm in Vietnam) The rainy season 2 1.501.50 to to1.99 wet -1.00toto -1.49 Moderate Moderate 1.99 Severe Severe wet -1.00 -1.49 drought drought is from April through October, representing 85-90% of the 1.49 Moderate wet -1.99 drought 3 1.001.00 to to1.49 Moderate wet -1.50 -1.50toto -1.99 SevereSevere drought total annual rainfall, and the dry season is from November to 0.99 Mild wet -2.00 or less Extreme drought to April representing only 10-15% of the total annual to 0.99 Mild wet -2.00 or less Extreme drought Streamflow Drought Index (SDI): similar to SPI, the SDI rainfall With regard to water availability, the river basin was developed explore the water resources conditions produces 136developed km3/year, oftowhich 83 km3 (61%) is generated StreamflowtoDrought Index (SDI): similar to SPI, the SDI was of thethe watershed based on conditions the information cumulativebased in Vietnam’s territory of explore water resources of theofwatershed on the information State SPI value Category SPI value Category cumulative streamflow volumes for reference periods, as expressed in equation [15]: ∑ (2) Vietnam Journal of Science, JUne 2019 • Vol.61 Number where Vi,k denotes the cumulative streamflow volume for the i-th hydrological year and the k-th reference period, k = for October-December, k = for October–March, k = Technology and Engineering 91 Environmental Sciences | Climatology Fig Map of the Red river basin (left) and the H-TBRW comprising five main tributaries in Vietnam (right) Precipitation and streamflow data reproducibility Due to transboundary issues, information about precipitation and streamflow in the portions beyond the border of Vietnam is not available to the public Attempts have been made to cover this problem through the provision of reanalysis products One of the recent precipitation products is APHRODITE Asian Precipitation Highly Resolved Observational Data Integration Towards the Evaluation - providing gridded daily precipitation over the Asia monsoon region from 1951 to 2015 APHRODITE has advantages for studies of water resources However, it is worth noting that APHRODITE is a reanalysis product based on historical measurement of precipitation, so it is not able to offer some type of quantitative projection in Cascading domains model of the downscaling model of the future In addition, featured 0.25-degree grid cells, Fig Fig with Cascading computational domains of computational the downscaling and location and location of observation sites in the H-TBRW (modified after APHRODITE is considered a coarse spatial resolution product observation sites in the H-TBRW (modified after [7]) [7]) that diminishes water resources studies at local scales With regard to streamflow data reproducibility, the downscaled precipitation is As a result, high spatialused and temporal resolution atmospheric that the simulated rainfall for the historical period to drive the Watershed Environmental Hydrology Model (WEHY) for1975-2006 hydrologic and streamflow data - which were already reconstructed and over the H-TBRW is comparable to the observed precipitation simulations in the H-TBRW The WEHY is a physically based hydrologic model that is verified for the entire Red river basin for period 1950-2015 [7, datasets either derived from direct point measurement or the on actual physical processes and information from the model 16] - are employed in this developed study to derivebased hydro-meteorological APHRODITE product Detailed model verification can be seen computational unit areas throughout the watershed domain [16] The model was also drought indices in the literature [7] designed for coupling regional climate models (e.g., the WRF model) through its land The high-resolution atmospheric and streamflow data are Withmodel regard parameters to streamflowaredata reproducibility, the surfacereproduced component addition, the nearly calibration-free a dynamic downscaling product using In a coupled downscaled precipitation is used to drive the Watershed because they are estimated on actual physical information of the catchment such regional hydroclimate model, or simply referred to as based the Environmental Hydrology Model (WEHY) for hydrologic as topography, and land use/cover Therefore, it illustrates advantages for the WEHY-HCM [7, 16] Atmospheric conditionssoil, were reproduced simulations in the H-TBRW The WEHY is a physically based using weather research forecast (WRF) of simulations The WRF in scattered observation catchments assessment water resources hydrologic model that is developed based on actual physical simulations were originally nested in the coarse resolution information from the model The which WEHY setupprocesses for the and H-TBRW was realized in computational the literatureunit [16] (1.25-degree) reanalysis data, ERA-20C, weremodel developed areas throughout the watershed domain [16] The model was For a short description, the entire H-TBRW was divided into computational units (or by the European Centre for Medium Range Weather Forecasts also designed for coupling regional climate models (e.g., the These simulations were performed for abased domain with a in topography and land surface information Runoff is sub-basins) on(D1) similarity WRF model) its land surface In addition,The spatial resolution of 81 km The WRFfrom simulations were then interaction generated the dynamic of through hillslope flow andcomponent channel routing the model parameters are nearly calibration-free because further refined through cascading domains of 27 km at (D2)Yen and 9Bai station were employed for model calibration monthly discharges and they are estimated based on actual physical information of km (D3), respectively, as illustrated in Fig The WRF provided validation Model performance statistics exhibited agreement between the monthly simulation outputs every three hours The simulated rainfall the catchment such as topography, soil, and land use/cover andscales observed discharges Nash it Sutcliffe Coefficients ofof0.87 illustratesEfficiency advantages for the assessment waterand was then aggregated into simulated larger temporal (e.g., daily or Therefore, 0.86 verification were obtained the model calibration andobservation validation, respectively Relative resources in scattered catchments monthly time series) for model Resultsfor illustrated errors in runoff volume were less than 5% These indicate a reasonable reproduction of the monthly discharges for the H-TBRW and useful application for further assessment of hydrologic conditions over the Red River basin Vietnam Journal of Science, JUne 2019 • Vol.61 Number 92 Technology and Engineering Results and discussion Fig Cascading computational domains of the downscaling model and loca observation sites in the H-TBRW (modified after [7]) With regard to streamflow data reproducibility, the downscaled precipit used to drive the Watershed Environmental Hydrology Model (WEHY) for hyd simulations in the H-TBRW The WEHY is a physically based hydrologic mode developed based on actual physical processes and information from the computational unit areas throughout the watershed domain [16] The model w designed for coupling regional climate models (e.g., the WRF model) through i surface component In addition, the model parameters are nearly calibrati because they are estimated based on actual physical information of the catchme as topography, soil, and land use/cover Therefore, it illustrates advantages assessment of water resources in scattered observation catchments The WEHY model setup for the H-TBRW was realized in the literatur For a short description, the entire H-TBRW was divided into computational un sub-basins) based on similarity in topography and land surface information Ru generated from the dynamic interaction of hillslope flow and channel routin monthly discharges at Yen Bai station were employed for model calibratio validation Model performance statistics exhibited agreement between the m simulated and observed discharges Nash Sutcliffe Efficiency Coefficients of 0.86 were obtained for the model calibration and validation, respectively R errors in runoff volume were less than 5% These indicate a reasonable reproduc the monthly discharges for the H-TBRW and useful application for further asse of hydrologic conditions over the Red River basin Results and discussion Climatological drought conditions over H-TBRW Climatological drought conditions over H-TBRW Environmental Sciences | Climatology The WEHY model setup for the H-TBRW was realized in for the remaining sub-catchments because rainfall remains the literature [16] For a short description, the entire H-TBRW an unpredictable variable among the others simulated by the was divided into computational units (or sub-basins) based WRF model It is understood as the uncertainties of the model on similarity in topography and land surface information structure, parameterization schemes, boundary, and initial Runoff is generated from the dynamic interaction of hillslope conditions In general, most model simulations tend to provide flow and channel routing The monthly discharges at Yen Bai information about a climatological trend rather than a precise station were employed for model calibration and validation simulation of an event magnitude and the time it occurs In Model performance statistics exhibited agreement between the addition, ground observation sites are quite scattered, leading monthly simulated and observed discharges Nash Sutcliffe to substantial errors for area rainfall estimates It is noted that Efficiency Coefficients of 0.87 and 0.86 were obtained for the calculated SPIs considering rainfall as a gamma distribution the model calibration and validation, respectively Relative variable outperform those calculated considering rainfall as a errors in runoff volume were less than 5% These indicate a normal distribution variable that tends to underestimate the reasonable reproduction of the monthly discharges for the drought conditions [17] H-TBRW and useful application for further 5.0 5.0 (B) Da river (A) Da river 5.0 5.0 assessment of hydrologic conditions over 5.0 5.0 4.0 4.0 (B) river (A) Da river 3-month 3-month (Rep) 1-month (Obs) 1-month (Rep) (B) Da Da river (Obs) (A) Da river 4.0 4.0 4.0 4.0 3.0 3.0 the Red river basin 3-month (Obs) 3-month (Rep) 1-month (Obs) 1-month (Rep) 3-month (Obs) 3-month (Rep) 1-month (Obs) 1-month (Rep) 3.0 3.0 This study first attempted to test the SPI derived from the reproduced precipitation using the Nash Sutcliffe Efficiency Coefficient, which can suggest the agreement in time and severity level of drought conditions of the SPI A test was conducted for the Da river sub-catchment over a period of five years (1990-1994) Results reveal that the simulated SPI and that obtained using observation data are quite similar, as seen in Fig Performance statistics are presented in Table and reveal encouraging results However, it is noted that similar SPI verification is quite challenging SPI SPI SPI SPI SPI SPI Time (month) (month) Time Time (month) 5.0 5.0 5.0 4.0 4.0 3.0 3.0 2.0 (C) Da river river (C) 6-month (C) Da river (Obs) 6-month (Obs) Time (month) (month) Time Time (month) 5.0 5.0 5.0 4.0 6-month (Rep) 4.0 3.0 6-month (Rep) 3.0 2.0 1.0 2.0 (D) Da river river (D) 9-month (D) Da river (Obs) 9-month (Obs) 9-month (Rep) SPISPI SPI 1.0 2.0 0.0 1.0 -1.0 0.0 -1.0 0.0 1.0 -1.0 0.0 -1.0 -2.0 -1.0 -2.0 -3.0 -2.0 -3.0 -2.0 -1.0 -2.0 -3.0 -2.0 -3.0 -4.0 -3.0 -4.0 -5.0 -4.0 -5.0 -4.0 -3.0 -4.0 -5.0 -4.0 -5.0 -5.0 -5.0 Time (month) Time (month) 5.0 5.0 4.0 5.0 4.0 3.0 4.0 3.0 2.0 3.0 2.0 1.0 2.0 1.0 0.0 1.0 0.0 -1.0 0.0 -1.0 -2.0 -1.0 -2.0 -3.0 9-month (Rep) Time (month) Time (month) Time (month) (E) Da river (E) Da river 12-month (Obs) (E) Da12-month river (Obs) 12-month (Rep) 12-month (Rep) 12-month (Obs) 12-month (Rep) 5.0 5.0 4.0 5.0 4.0 3.0 4.0 3.0 2.0 3.0 2.0 1.0 2.0 1.0 0.0 1.0 0.0 -1.0 0.0 -1.0 -2.0 -1.0 -2.0 -3.0 Time (month) (F) Da river (F) Da river 24-month (Obs) 24-month (Rep) 24-month (Rep) (F) Da24-month river (Obs) 24-month (Obs) 24-month (Rep) SPISPISPI In general, droughts last from a couple of months to a few years This study attempts to understand climatological drought conditions corresponding to the time scales of one, three, six, nine, 12, and 24 months The previous study [7] revealed a reasonable agreement of the reproduced monthly precipitation over the H-TBRW with the APHRODITE product However, this study again performs the verification of SPI derived from the reproduced precipitation data against those determined using raingauge measurements The verification is conducted on a sub-basin average basis As illustrated in Fig 1, the H-TBRW is delineated into five sub-catchments, namely, Da, Thao, Lo-Gam, and Upper Thai Binh sub-catchments, and the Red river delta Available observed precipitation data during the period from 1975 to 2006 are employed for the SPI verification SPISPI SPI Climatological drought conditions over H-TBRW 3.0 2.0 2.0 2.0 1.0 1.0 1.0 0.0 0.0 0.0 -1.0 -1.0 -1.0 -2.0 -2.0 -2.0 -3.0 -3.0 -3.0 -4.0 -4.0 -4.0 -5.0 -5.0 -5.0 3.0 2.0 2.0 2.0 1.0 1.0 1.0 0.0 0.0 0.0 -1.0 -1.0 -1.0 -2.0 -2.0 -2.0 -3.0 -3.0 -3.0 -4.0 -4.0 -4.0 -5.0 -5.0 -5.0 SPISPISPI Results and discussion -2.0 -3.0 -4.0 -3.0 -4.0 -5.0 -2.0 -3.0 -4.0 -3.0 -4.0 -5.0 -4.0 -5.0 -4.0 -5.0 -5.0 -5.0 Time (month) Time (month) Time (month) Time (month) Verification of(month) SPIs Time Fig the Da River sub-catchment Time with (month)different time Fig Verification of SPIs forforthe Da river sub-catchment with different time Fig (a) Verification for the River sub-catchment different time scales: 1-month; of (b)SPIs 3-month; (c)Da 6-month; (d) 9-month; (e)with 12-month; and (f) scales: (a) 1-month; (b) 3-month; (c) 6-month; (d) 9-month; (e) 12-month; and Fig (a) Verification of SPIs for the Da River sub-catchment with different time scales: 1-month; (b) 3-month; (c) 6-month; (d) 9-month; (e) 12-month; and (f) 24-month (f)scales: 24-month (a) 1-month; (b) 3-month; (c) 6-month; (d) 9-month; (e) 12-month; and (f) 24-month Thus, the next attempts are focusing on model verification in terms of 24-month Thus, drought the nexttrend attempts are For focusing onFigs model verification of climatological and risk example, and illustrateintheterms drought Table Statistics of next SPI verification for focusing Da sub-catchment Thus, the attempts are on model verification in terms of climatological drought trend and and risk For example, Figs and illustrate the drought trends (time scales of three six months) obtained from model simulation versus climatologicalNash drought trend and risk For example, Figs and illustrate the drought Sub-catchment Efficiency Coefficient trends (time scalesSutcliffe of three and six months) obtained from model simulation versus8 trends (time scales and six months) versus8 1-month of three 3-month 6-month obtained 9-monthfrom model 12-monthsimulation 24-month Da river 0.37 0.58 0.65 0.69 0.56 NA NA: not applicable JUne 2019 • Vol.61 Number Vietnam Journal of Science, Technology and Engineering 93 SP found for the remaining time scales Table demonstrates the risk of severe drought -1.0 conditions (represented by a number of drought events that SPI is less than minus-1.5) in -2.0 the Red River Delta and Upper Thai Binh sub-catchment These results indicate a -3.0 reasonable performance of the model simulation against observation On average, the Environmental Sciences | Climatology number of severe drought events are well reproduced by the WRF model However, the -4.0 Time (month) model still provides the average absolute relative errors of about 20% 3.0 2.0 3.0 (A) Red river delta 3-month (Obs) 3-month (Rep) 2.0 3.0 1.0 2.0 0.0 1.0 1.0 6-month (Obs) 6-month (Rep) (A) Upper Thai Binh river 3-month (Obs) 3-month (Rep) SPI SPI 0.0 (B) Red river delta -1.0 0.0 SPI -1.0 -2.0 -1.0 -2.0 -3.0 -3.0 -2.0 -4.0 -4.0 -3.0 -4.0 Time (month) Time (month) Fig Verification of SPI trend for the Red river delta with time scales: (a) 3-month; (b) 6-month Time (month) Fig Verification of SPI trend for the Red River Delta with time scales: (a) 3.0 3.0 2.0 2.0 1.0 (B) Red river delta (A) Upper Thai Binh river 6-month (Obs) 6-month (Rep) 3-month (Obs) 3-month (Rep) 3.0 3-month; (b) 6-month 2.0 (B) Upper Thai Binh river 6-month (Obs) 6-month (Rep) 1.0 SPISPI 1.0 0.0 0.0 -1.0 -1.0 -2.0 -2.0 -3.0 -3.0 -4.0 -4.0 SPI 0.0 -1.0 -2.0 -3.0 -4.0 Time (month) Time (month) Time (month) for the Delta with timeriver scales: (a) Fig Verification of SPI trendtrend Fig.4.3.05 Verification of SPI forRed theRiver Upper Thai Binh sub-catchment with time scales: (a) 3-month; (b) 6-month (B) Upper Thai Binh river 6-month (Obs) 6-month (Rep) Fig Verification of SPI trend for the Upper Thai Binh River sub-catchment SPI 3-month; 2.0 (b) 6-month Table Number of severe drought events simulated by model 1.0 versus actual observation during 1975-2006 in Red river delta and0.0Upper Thai Binh river sub-catchment -1.0 Sub-catchment -2.0 1-month 3-month 6-month 9-month 12-month 24-month Average 27 21 26 14 Red river delta -3.0 Observation 20 19 21 with time scales: (a) 3-month; (b) 6-month Thai Binh sub-catchment These results indicate a reasonable performance the model against observation Table Number of of severe drought simulation events simulated by model versus actual On average, the number of severe drought events are well observation during 1975-2006 in Red River Delta and Upper Thai Binh River reproduced by the WRF model However, the model still provides the average absolute relative errors of about 20% sub-catchment Sub-catchment 1-month 3-month 6-month 9-month 12-month 24-month Average climatological drought conditions of various -4.0 Model 35 23 21 16 14 17 21 time scales in the Observation 27 H-TBRW 20 21are reproduced 19 26 based 14 on the 21 Absolute relative error (%) 30% 15% 0% 16% 46% 21% 21% simulated rainfall for the period 1950-2015, a sufficiently Model 35 23 21 16 14 17 21 Time (month) long relative timeerrorscale that is15%able 0%to reflect the accurate Upper Thai Binh river Absolute (%) 30% 16% 46% most21% 21% climatological Upper Thai Binh River condition in comparison with such studies as Observation 27 20 21 19 26 14 21 Fig Verification of SPI trend for the Upper Thai Binh River sub-catchment Observation 27 20the drought 21 19 26 14 shorter 21 [18, 19], which assessed conditions using Model 35 23 6-month 21 16 14 17 21 with time scales: (a) 3-month; (b) Model 35 illustrates 23 21 an example 16 14 17 21 periods of time Fig of climatological Absolute relative error (%) 30% 15% 0% 16% 46% 21% 21% Table Number of severe drought events simulated by model versus actual Absolute relativeconditions error (%) 30% with 15% 0% scale 16% of six 46% months 21% in the 21% drought the time observation during 1975-2006 in Red River Delta and Upper Thai Binh River H-TBRW Results show there has been a slight increase of 10 Thus, the next attempts are focusing on model sub-catchment verification in terms of climatological drought trend and drought conditions in the Red river delta, Lo-Gam, and Thai Sub-catchment 1-month 9-month the 12-month 24-month Average Binh sub-watersheds; meanwhile, an intensified implication risk For example, Figs.3-month and6-month illustrate drought trends Red River Delta (time scales of three and six months) obtained from model of drought has been observed for Da and Thao subwatersheds It is not revealed in this text; however, in terms Observation 21 19 26 14 21 simulation versus27actual20observation during 1975-2006 in the of time scales, the drought conditions have been more severe Model 35 Upper 23 Thai 21 Binh16sub-catchment 14 17 Results 21 Red river delta and with increased time scales Table presents the number of Absolute relative error (%) 15% 0% and observed 16% 46% drought 21% trends 21% demonstrate that30% both modelled Upper Binh River are Thai comparable and indicate a slight decrease in droughts climatological severe and extreme droughts in the H-TBRW during 1950-2015 Among the five sub-watersheds, the Red Observation 27 20 21 19 26 the remaining 14 21 Similar results (not shown) are found for Model 35 demonstrates 23 21 14 severe 17 drought 21 river delta and Upper Thai Binh sub-watershed experienced time scales Table the16 risk of Absolute relative error (%) 30% 15% 0% 16% 46% 21% 21% conditions (represented by a number of drought events that more severe drought events; however, the Da sub-watershed SPI is less than minus-1.5) in the Red river delta and Upper 10 has observed more extreme drought events 94 Vietnam Journal of Science, Technology and Engineering AsDelta a result, Red River JUne 2019 • Vol.61 Number (C) Thao river 6-month (Rep) 2.0 and Thai Binh sub-watersheds; meanwhile, an intensified implication of drought has 1.0 SPISPI been observed for Da and Thao sub-watersheds It is not revealed in this text;1.0 however, 0.0 0.0 in terms of time scales, the drought conditions have been more severe with-1.0increased -1.0 -2.0 time scales Table presents the number of climatological severe and extreme droughts -2.0 y = -3E-06x + 0.1033 y = -2E-05x + 0.4774 -3.0 Fig.the6.H-TBRW Climatological conditions simulated by model during 1950-2015 in during drought 1950-2015 Among the five sub-watersheds, the Red River -3.0 -4.0 for theand H-TBRW sub-catchment: (a) Red experienced River Delta,more (d) Dasevere River,drought (c) -4.0 Thao River, Delta Upper Thai Binh sub-watershed events; (d) Lo-Gam River, and (e) Upper Thai Binh River with the straight lines (d) Lo-Gam and (e)has Upper Thaimore Binh Riverdrought with events the straight lines however, the DaRiver, sub-watershed observed extreme Time (month) Time (month) representing the trend of the drought conditions (A) Red river delta (C) Thao river 3.0 3.0 6-month (Rep) 6-month (Rep) 3.0 6-month (Rep) 6-month (Rep) SPI SPI SPI -2.0 -2.0 -2.0 y = -2E-05x + 0.496+ 0.1315 y = -4E-06x -3.0 -3.0 -4.0 Time (month) Time Time(month) (month) 3.0 (B)Lo-Gam Da riverriver (D) 3.0 6-month (Rep) 6-month (Rep) 2.0 1.0 (E) Upper Thai Binh river 6-month (Rep) SPI SPI -1.0 0.0 -1.0 -2.0 y = -2E-05x + 0.496 + 0.1315 y = -4E-06x y = -3E-06x + 0.099 -3.0 -4.0 -4.0 Time(month) (month) Time Time (month) 3.0 Fig (cont’d) SPI (E) Upper Thai Binh river 6-month (Rep) Fig Climatological drought conditions simulated by model during 1950-2015 for the 2.0 11 H-TBRW sub-catchment: (a) Red river delta, (B) Da river, (c) Thao river, (d) Lo-Gam 1.0 river, and (e) Upper Thai Binh river with the straight lines representing the trend of 0.0 the drought conditions -1.0 -2.0 Table Climatological severe and extreme drought events y = -3E-06x + 0.099 -3.0 simulated by model during 1950-2015 in Red river delta and Upper Thai Binh river sub-catchment -4.0 1-month 3-month 6-month 9-month 12-month 24-month Average Time (month) 3.0 2.0 Red river delta (A) Hoa Binh 65 55 53 50 45 54 54 Extreme drought 30 21 15 10 27 19 Severe drought 54 53 56 48 44 49 51 Extreme drought 20 29 26 25 21 19 23 12 Time (month) Fig (cont’d) -2.0 51 50 49 48 37 49 Extreme drought 22 12 15 23 23 14 18 Severe drought 45 50 57 59 59 45 53 Extreme drought 18 16 13 12 12 13 Time (month) 2.0 74 58 56 48 47 38 54 22 21 19 23 20 20 21 October-March (Rep) SPI SPI 0.0 -1.0 -2.0 Upper Thai Binh river Extreme drought (B) Yen Bai 1.0 Lo-Gam river Severe drought 12 -4.0 3.0 57 y = -1E-05x + 0.4359 -3.0 Thao river Severe drought October-March (Rep) -4.0 0.0 -1.0 Fig (cont’d) Severe drought Da river 6-month (Rep) y = -3E-06x + 0.099 -3.0 1.0 SPI SPI Sub-catchment (D) Lo-Gam river 2.0 the drought situation in the Thao 11 1.0 river is rather stable The drought 0.0 situations (not shown) in other rivers -1.0 are also found to be similar These -2.0 trends indicate minor stress on water y = -4E-06x + 0.1315 -3.0 availability for the water-use sectors -4.0 in the downstream areas However, it is noted that the influence of reservoir Time (month) operation is excluded from the 3.0 (E) Upper Thai Binh river 6-month (Rep) streamflow simulations Thus, the next 2.0 effort of this research series will further 1.0 elaborate this trend of drought as both 0.0 reservoir operation and projection data 12 -1.0 are analyzed -2.0 SPI 1.0 0.0 -3.0 y = -2E-05x + 0.4774 -3.0 -4.0 -2.0 0.0 -1.0 -1.0 -1.0 SPI 2.0 6-month (Rep) 1.0 0.0 0.0 y = -3E-06x + 0.1033 y = -2E-05x + 0.4774 (C) Thao river 2.0 1.0 1.0 Time (month) Time (month) 3.0 representing the trend of the drought conditions (B)Lo-Gam Da riverriver (D) 2.0 2.0 SPISPI 3.0 3.0 2.0 2.0 1.0 1.0 0.0 0.0 -1.0 -1.0 -2.0 -2.0 -3.0 -3.0 -4.0 -4.0 Fig Climatological drought conditions simulated by model during 1950-2015 | Climatology Environmental for the H-TBRW Sciences sub-catchment: (a) Red River Delta, (d) Da River, (c) Thao River, -3.0 y = 4E-07x - 0.0026 -4.0 Time (month) Hydrological drought conditions over the H-TBRW The present study examines hydrological drought conditions based on the reproduced streamflow at various sites in the H-TBRW The hydrological drought conditions are explored for different time periods Within this text, Fig illustrates the hydrological drought trends over the past 65 years at the Da and Thao rivers (Hoa Binh and Yen Bai, respectively) It appears that the hydrological drought in the Da river is becoming slightly severe; meanwhile, Fig.Fig Hydrological drought conditions simulated by model Hydrological drought (October-March) conditions (October-March) during 1950-2015 the H-TBRW sub-catchment: Da H-TBRW River (at Hoa simulated by for model during 1950-2015 for(a)the sub-Binh), andcatchment: (b) Thao River (at Yen Bai) representing the trend of the drought conditions (a) Da river (at Hoa Binh), and (b) Thao river (at Yen Bai)and representing the trend of the drought conditions Conclusion remarks Understating hydro-climatological conditions in a transboundary is always Conclusions and remarks challenging because of the insufficient data availability The present study has explored the hydro-climatological conditions over the H-TBRW based in on the Understating drought hydro-climatological conditions downscaled rainfall and reproduced streamflow by the state-of-the-art WEHY-HCM a transboundary is always challenging because of the model The resultsdata demonstrate a slight The increase in trends of both insufficient availability present study hasclimatological explored and hydrological conditions (SPI and SDI) Over the H-TBRW, the Da and Thao rivers are expecting a stronger implication of drought; meanwhile, the remaining rivers are quite likely to experience similar drought conditions as in the past It is also noted that there exist modelJournal intrinsic uncertainties because of Vietnam of Science, imperfect model structure, parameterization schemes, boundary, and initial conditions Technology and Engineering In general, model simulations provide reasonable climatological trends rather than a precise simulation of an event magnitude and the time it occurs As a result, model bias JUne 2019model • Vol.61 Number imperfect structure, parameterization schemes, boundary, and initial95 conditions Environmental Sciences | Climatology the hydro-climatological drought conditions over the H-TBRW based on the downscaled rainfall and reproduced streamflow by the state-of-the-art WEHY-HCM model The results demonstrate a slight increase in trends of both climatological and hydrological conditions (SPI and SDI) Over the H-TBRW, the Da and Thao rivers are expecting a stronger implication of drought; meanwhile, the remaining rivers are quite likely to experience similar drought conditions as in the past It is also noted that there exist model intrinsic uncertainties because of imperfect model structure, parameterization schemes, boundary, and initial conditions In general, model simulations provide reasonable climatological trends rather than a precise simulation of an event magnitude and the time it occurs As a result, model bias correction will be still needed for further interpretation of the hydro-climatological drought conditions in such sub-catchments of the H-TBRW ACKNOWLEDGEMENTS This study was financially supported by the National Science and Technology Program for 2016-2020 (KC.08.05/16-20), Ministry of Science and Technology in Vietnam The study was implemented at the Key Laboratory of River and Coastal Engineering (KOLRCE-Vietnam) The authors declare that there is no conflict of interest regarding the publication of this article REFERENCES [1] K.N Kumar, et al (2013), “On the observed variability of monsoon droughts over India”, Weather Clim Extrem., 1, pp.42-50 [2] IPCC (2018), Climate Change 2013: The Physical Science Basis Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK and New York, NY, USA, 1585pp [3] K Takara, et al (2009), “Assessing climate change impact on water resources in the Tone river basin, Japan, using super-high resolution atmospheric model output”, J Disaster Res., 4(1), pp.12-22 [4] D.H Burn, M.A.H Elnur (2002), “Detection of hydrologic trends and variability”, J Hydrol., 255, pp.107-122, http://dx.doi.org/10.1016/ S0022-1694(01)00514-5 [5] L.H Benjamin, M.A Saunders (2002), “A drought climatology for Europe”, International Journal of Climatology, 22, pp.1571-1592 [6] A Dai (2013), “Increasing drought under global warming in observations and models”, Nat Clim Change, 3, pp.52-58, http://dx.doi org/10.1038/nclimate1633 [7] C Ho, A Nguyen, A Ercan, M.L Kavvas, V Nguyen, T Nguyen 96 Vietnam Journal of Science, Technology and Engineering (2018), “Assessment of atmospheric conditions over the Hong-Thai Binh river watershed by means of dynamically downscaled ERA-20C reanalysis data”, Journal of Water and Climate Change, jwc2018291, Doi: 10.2166/wcc.2018.291 [8] W.C Palmer (1965), Research Paper No 45 - Meteorological drought, U.S Department of Commerce, Weather Bureau, Washington D.C [9] W.C Palmer (1968), “Keeping track of crop moisture conditions, nationwide: the new Crop Moisture Index”, Weatherwise, 21(4), pp.156161 [10] B.A Shafer, L.E Dezman (1982), “Development of a Surface Water Supply Index (SWSI) to assess the severity of drought conditions in snowpack runoff areas”, Proceedings of the Western Snow Conference, pp.164-175, Colorado State University, Fort Collins, Colorado [11] R.A Seiler, M Hayes, L Bressan (2002), “Using the standardized precipitation index for flood risk monitoring”, International Journal of Climatology, 22(11), pp.1365-1376 [12] S.K Min, W.T Kwon, E.H Park, Y.G Choi (2003), “Spatial and temporal comparisons of droughts over Korea with East Asia”, International Journal of Climatology, 23, pp.223-233 [13] S Morid, V Smakhtin, M Moghaddasi (2006), “Comparison of seven meteorological indices for drought monitoring in Iran”, International Journal of Climatology, 26(7), pp.971-985 [14] National Centers for Environmental Information, https://www ncdc.noaa.gov/sotc/drought/201904#det-spi [15] I Nalbantis, G Tsakiris (2009), “Assessment of hydrological drought revisited”, Water Resour Manag., 23(5), pp.881-897 [16] C Ho, T Trinh, A Nguyen, Q Nguyen, A Ercan, M.L Kavvas (2019), “Reconstruction and evaluation of changes in hydrologic conditions over a transboundary region by a regional climate model coupled with a physically-based hydrology model: application to Thao river watershed”, Science of The Total Environment., 668, pp.768-779 [17] H.V Cuong, N.T.N Nhan, T.V Bach, T.Q Toan (2019), “Research on drought conditions in the Red - Thai Binh basin with meteorological and hydrological data recovered from the combined model WEHY-WRF”, Journal of Water Resources Science and Technology, 52, pp.48-64 (in Vietnamese) [18] N.V Thang, et al (2015), National Science and Technology Program, code KC.08.17/11-15 Study and establish a drought forecast and warning system for Vietnam with a duration of up to months (in Vietnamese) [19] V.T Hang, T.T Ha (2013), “Comparison of drought indices for climatological regions of Vietnam”, VNU Journal of Science: Natural Sciences and Technology, 2S(29), pp.51-57 (in Vietnamese) JUne 2019 • Vol.61 Number ... H-TBRW River (at Hoa simulated by for model during 1950-2015 for( a)the sub -Binh) , andcatchment: (b) Thao River (at Yen Bai) representing the trend of the drought conditions (a) Da river (at Hoa Binh) ,... StreamflowtoDrought Index (SDI): similar to SPI, the SDI was of thethe watershed based on conditions the information cumulativebased in Vietnam s territory of explore water resources of theofwatershed... Forecasts also designed for coupling regional climate models (e.g., the These simulations were performed for abased domain with a in topography and land surface information Runoff is sub-basins)