development and application of improved long...

Research ArticleDevelopment and Application of Improved Long-TermDatasets of Surface Hydrology for Texas

Kyungtae Lee1 Huilin Gao1 Maoyi Huang2 Justin Sheffield34 and Xiaogang Shi5

1Department of Civil Engineering Texas AampM University College Station TX 77843 USA2Earth System Analysis and Modeling Group Pacific Northwest National Laboratory Richland WA 99352 USA3Department of Civil and Environmental Engineering Princeton University Princeton NJ 08544 USA4Geography and Environment University of Southampton Southampton SO17 1BJ UK5Department of Civil Engineering Xirsquoan Jiaotong-Liverpool University Suzhou 215123 China

Correspondence should be addressed to Huilin Gao hgaociviltamuedu

Received 5 October 2016 Revised 31 January 2017 Accepted 8 February 2017 Published 6 March 2017

Academic Editor Olga Zolina

Copyright copy 2017 Kyungtae Lee et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

Freshwater availability and agricultural production are key factors for sustaining the fast growing population and economy in thestate of Texas which is the third largest state in terms of agricultural production in the United States This paper describes a long-term (1918ndash2011) grid-based (18∘) surface hydrological dataset for Texas at a daily time step based on simulations from the VariableInfiltration Capacity (VIC) hydrological model The model was calibrated and validated against observed streamflow over 10 Texasriver basins The simulated soil moisture was also evaluated using in situ observations Results suggest that there is a decreasingtrend in precipitation and an increasing trend in temperature in most of the basins Droughts and floods were reconstructed andanalyzed In particular the spatially distributed severity and duration of major Texas droughts were compared to identify newcharacteristicsThemodeled flood recurrence interval and the return period were also compared with observations Results suggestthe performance of extreme flood simulations needs further improvement This dataset is expected to serve as a benchmark whichmay contribute to water resources management and to mitigating agricultural drought especially in the context of understandingthe effects of climate change on crop yield in Texas

1 Introduction

Texas the largest state in the contiguous United States(CONUS) contains awide range of climate regimes from aridto subtropical humid [1] The diverse climate range in Texasmanifests itself as large spatial and temporal variations inprecipitation and temperature Due to these large spatial andtemporal variations of rainfall and temperature hydrologicextreme events (such as droughts and floods) have led toadverse conditions for agricultural production [2] This is apressing issue for Texas which has the largest farm area andthe highest livestock production among the 50 states OverallTexas ranks third with regard to agricultural production [3]

During the past century Texas has experienced a numberofmajor drought and flood events [4ndash7] Among theweather-related disasters drought ranks first in causing loss of lifeand second in causing property loss [5] Drought in the

United States results in an estimated average annual damageof between 6 and 8 billion dollars [8 9] As a slow-motiondisaster drought brings a series of calamities to Texas lifeincluding dust storms crop failures livestock losses andeconomic crises The recent record drought in 2011 left thestate with 76 billion dollars in agricultural losses and with amultitude of dried up lakes and rivers [10] Unlike droughtswhich persist for months and longer floods are usuallytriggered by heavy rainfall during a short period of time[11] Flooding depends on a number of factors such as themagnitude and intensity of rainfall antecedent soil moistureconditions topography of the affected landscape soil typeand land use [12 13] Between 1985 and 2014 flooding causedan average of 82 deaths and $79 billion in property damageannually across the US [14] With 840 lives lost between 1959and 2008 Texas has the highest incidence of flood relatedfatalities among all 50 states [15] Texas is also the only state

HindawiAdvances in MeteorologyVolume 2017 Article ID 8485130 13 pageshttpsdoiorg10115520178485130

2 Advances in Meteorology

that has reported flood related fatalities in every single yearduring that same period [16]

While battling these extreme events Texas has becomea water deficient state where the demands for fresh waterhave been exacerbated by a rapidly growing populationThesewater issues are further challenged by climatic and land usechanges both of which may alter the natural hydrologicprocesses With a changing climate hydrologic extremes areprojected to become more frequent more severe and moreuncertain [17ndash19] Additionally the increasing portion ofimpervious land cover (due to urbanization) has a directeffect on elevating flood peaks [20] Due to the importanceof water resources for Texas and its vulnerability to water-related extreme events it is necessary to understand howfuture changes may impact Texasrsquo water resources and (riversystem) water budgets [21]

In this context comprehensive and reliable hydrologicdatasets which can support the analysis of historical hydro-logic extreme events are essential Specifically high qualitydatasets can be used to identify the onset and demiseof droughts and floods along with the multiple feedbackprocesses associatedwith hydrological extremes [11] Further-more such datasets can serve as a benchmark to evaluatefuture extreme events and to prevent record setting disastersin advance (through combining effective water resourcesmanagement measures with model predictions)

With the enhanced computational capabilities a highvolume of hydrological datasets have recently been generated(and released) for studying droughts and floods For instancethe North American Land Data Assimilation System-2(NLDAS-2 [22]) includes long-term (1979ndashpresent) simu-lations of the surface hydrology for the contiguous UnitedStates at 18∘ resolution This dataset has been used forproviding long-term records of water budget terms foranalyzing historic droughts and for providing the basis forseasonal drought prediction [23ndash25] These types of datasetshave also been widely used in regional assessments of climatechange impacts on surface hydrology such as in a set ofstudies focused on the Colorado River basin [26 27] Suchmodeled hydrologic datasets have strong advantages overtraditional observation based datasets whose availability islimited in time and space For example in situ observationsare often available only at point locations or over areas muchsmaller than the model spatial resolution Comparisons aretherefore restricted to the temporal and spatial scales resolvedby the model [28]

Although many of the above-mentioned hydrologicdatasets contain gridded long-termmodeled results over theentire state of Texas the data quality is often inadequate tosupport decisionmaking Typically only a very small numberof the Texas river basins (only one or two of them) havebeen calibrated against observed streamflow (eg [28 29])The results of a study by Oubeidillah et al [30] whichcalibrated for 2107 hydrologic subbasins (8-digit hydrologicunits HUC8s) over the entire CONUS show that the Nash-Sutcliffe values for most Texas basins are negative Withouta well-tested reliable dataset all analyses will be at riskfor providing misleading conclusions and recommendationsTherefore there is a strong need for effectively constraining

the quality of hydrologic model simulation results throughcalibration over each individual river basin in Texas

Driven by the meteorological forcings of Livneh et al([29] hereafter L13) we hereby provide a calibrated andvalidated hydrological dataset for 10 major Texas river basinsThe dataset is deemed high quality because of its relativelyhigh spatial (18∘) and temporal (daily) resolutions and itsevaluated skill compared to observed hydrological variablesThe dataset includes evapotranspiration runoff and soilmoisture records from 1918 to 2011 The L13 meteorologicalforcings are available from 1915 to 2011 but we used the firstthree years for model spin-up (and then analyzed from 1918onward) The dataset generated from this study was utilizedto fulfill two research objectives (1) to evaluate the impactsof a changing climate on the water budget terms and (2)to reveal new perspectives about hydrologic extreme events(such as droughts and floods) which cannot be assessed usingtraditional observations This study is organized as followsthe data and methods are presented in Section 2 where thecalibration of the soil parameters and the validation of thesimulated results are described The quality of the simulatedsoil moisture is evaluated against observed soil moisture InSection 3 the differences in water budget terms are studiedby comparing two periods 1918ndash1959 (Period 1) and 1960ndash2011 (Period 2) Historical drought (severity and duration)and flood (recurrence interval and return period) events areinvestigated based on simulated hydrologic variablesWe alsocompare the annual cycle of the water budget at each riverbasin between the two historical periods Finally discussionand conclusions are presented in Section 4

2 Data and Methodology

21 Study Area This study focuses on the Texas Gulf Regionlocated in southern central North America (25ndash34∘N 93ndash103∘W) which has a total area of 343100 km2 The regionincludes 10 major river basins (Figure 1) and covers fiveclimate zones (from arid to subtropical humid) The domaincontains geographical properties varying from dessert (farwest Texas) to mountainous (Guadalupe Range) regions [31]The diverse climate in Texas manifests itself as large spatialand temporal variations in precipitation and temperatureThe annual mean precipitation in Southeast Texas is morethan 1400mm while Northwest Texas only receives about400mm [32] The annual mean temperature varies greatlywith latitude from north to south According to Bomar [33]the average annual temperature (1961ndash1990) in the northernportion of the Texas High Plains is 132∘C while it is 233∘Cin Southern TexasWhile most west Texas rivers flow for onlypart of the year (due to a lack of precipitation) East Texasrivers flow year-round benefiting from a subtropical climate[34]

With an annual economic revenue of $100 billion agri-culture is very important in Texas A total of 528000 km2 isoccupied by farms and ranches About 76 of Texas surfacearea is occupied by farms and ranches and 22 of this is cropland For the crop land portion about 57 is harvested 10 isgrassland and 33 is either not harvested or fails to produce

Advances in Meteorology 3

Soil moisture observation sitesMajor rivers

0 125 250 500 750 1000(Kilometers)

Figure 1 Location of the ten river basins and soil moistureobservation sites used in this study

crops [35] The soil types in Texas range from clay to sandwith more than 1300 different varieties of soil

22 VIC Model A semidistributed macro scale hydrologicalmodel the Variable Infiltration Capacity (VIC) model [36]was used to generate the long-term hydrologic budget inthis study The VIC model has been widely utilized forassessing water resources land-atmosphere interactions andthe overall hydrological budget (and its responses to weatherand climate) over many river basins around the world [2837ndash41] In the Jinghe basin located in Northwest Chinaan assessment of the river system changes under both achanging climate and human activities was implementedusing VICmodeled streamflow [42]TheVICmodel was alsoemployed to generate a forecast of soil moisture runoff andstreamflow for the Yellow River in China [43] VIC simulatedsoil moisture and runoff have made significant contributionsto drought studies [44ndash49] The VIC model has been welladopted for continental to global scale drought monitoringand forecasting using soil moisture and streamflow [5051] Soil moisture on one hand is a critical variable forquantifying drought severity and extent but on the otherhand it is typically not observed on a large scale over along periodTherefore soil moisture simulated by hydrologicmodelsmdashsuch as the VIC modelmdashmay serve as the bestalternative (to observations) at regional to global scales [2846 52 53]

The VIC model parameters can be classified into twogroups those that are prescribed and those that are calibratedIn this study the soil and vegetation parameters that do not

require calibration were adopted from the NLDAS-2 param-eters at 18∘ resolution [54] The model was used to simulatethe water and energy budgets with the major hydrologicflux terms (eg evapotranspiration) and the state variables(eg soil moisture) simulated at a daily time step The VICmodeled surface runoff and base flow at each grid cell werethen routed through a Digital Elevation Model (DEM) basedriver network to generate streamflow estimations for eachbasin [55]

23 Meteorological Forcings The observation based meteo-rological daily forcings from 1915 to 2011 were adopted fromthe L13 dataset to drive the VIC model The grid-based L13dataset includes four meteorological variables precipitationwind speed and daily minimum and maximum tempera-ture The precipitation and temperature observations wereprovided by National Climatic Data Center (NCDC) andCooperative Observer (COOP) stations The SynergraphicMapping System (SYMAP) algorithm [56] was employedto generate the gridded temperature and precipitation at116∘ resolution from the point dataThe Parameter-ElevationRegressions on Independent Slopes Model (PRISM) wasthen used to match the long-term mean of the griddedprecipitation data which was scaled on a monthly basis [57]Wind speed values obtained from the National Centers forEnvironmental Prediction-National Center for AtmosphericResearch (NCEP-NCAR) reanalysis [58] were linearly inter-polated from 19∘ resolution (approximately) to 18∘ reso-lution To match the spatial resolution of the VIC modelparameters the 116∘ forcings from L13 were rescaled up to18∘ using the nearest-neighbor interpolation method

24 Model Calibration An automated optimization tech-nique Multiobjective Complex evolution (MOCOM-UA[59]) was employed to calibrate the VIC model over the 10major rivers in Texas During the calibration process themean absolute error (MAE) and the Nash-Sutcliff coefficient[60] were used as the objective functions to minimize thedifference between the simulated and observed streamflowThe monthly streamflow observations at the US GeologicalSurvey (USGS) stations closest to the river outlets were usedfor both calibration and validation purposes (Table 1)

The calibration aimed to find the best soil parametervalues for minimizing the difference between observed andsimulated monthly streamflow over the calibration period(1960ndash1985) Six VIC soil parameters were selected forcalibration based on sensitivity analysis [61] including thevariable infiltration curve parameter (119887inf ) the exponent ofthe Brooks-Corey drainage equation (exp) the thickness ofsoil layers 2 and 3 (1198632 and1198633) the fraction of the maximumvelocity of base flow at which nonlinear base flow begins(119863119904) and the fraction ofmaximum soilmoisture abovewhichnonlinear base flow occurs (119882119904) The calibration involvessetting an identical soil parameter set for each basin to findthe best combination of the six parameters Although thecalibration period is about 03∘C cooler than the annualtemperature over the entire period sensitivity test results (notshown) suggest that the temperature impacts on streamflow


Table 1 Ten river basins and streamflow gauge stations

Name Abbreviation USGS station Latitude (N) Longitude (W) Basin area (km2) PeriodSabine SABIN 08030500 30∘1810158401310158401015840 93∘4410158403710158401015840 19617 1924ndash2016Neches NECHE 08041000 30∘2110158402010158401015840 94∘0510158403510158401015840 25752 1922ndash2016Trinity TRNTY 08066250 30∘3410158401910158401015840 94∘5610158405510158401015840 46418 1965ndash2016Brazos BRAZO 08111500 30∘0710158404410158401015840 96∘1110158401510158401015840 111077 1938ndash2016Colorado COLOR 08162000 29∘1810158403210158401015840 96∘0610158401310158401015840 102172 1938ndash2016Guadalupe GUADA 08175800 29∘0510158402510158401015840 97∘1910158404610158401015840 15426 1964ndash2016San Antonio SANAN 08188500 28∘3810158405710158401015840 97∘2310158400510158401015840 10831 1924ndash2016Nueces NUECE 08211000 28∘0210158401710158401015840 97∘5110158403610158401015840 43276 1939ndash2016San Jacinto SANJA 08068000 30∘1410158404010158401015840 95∘2710158402510158401015840 10199 1924ndash2016Lavaca LAVAC 08164000 28∘5710158403510158401015840 96∘4110158401010158401015840 5985 1938ndash2016

J F MA M J J A S ON D0

200

400

600

J F MAM J J A S O ND0

200

400

600

J F MAM J J A S OND0

200

400

600


200

400

600


200

400

600


50

100

150

J FMAM J J A S OND0

50

100

150


50

100

150

J FMAM J J A S O ND0

20

40

60

J F MAM J J A S O N D0

20

40

60

SABIN

Month Month Month Month Month


OBSL13SIM

NECHE TRNTY BRAZO COLOR

GUADA SANAN NUECE SANJA LAVAC

(m3 s

)(m

3 s)

Figure 2 Monthly observed (OBS) L13 and calibrated (SIM) streamflow (1960ndash1985)

in Texas river basins are ignorable This can be explainedby the fact that Texas is water limited and rainfall typicallyoccurs at large rates over short periodsmdashwhich makes soilmoisture and streamflow insensitive to small variations ofannual temperature

Figure 2 compares the annual cycle of the calibratedmonthly streamflow with observations over the 10 majorTexas river basins and Table 2 lists the statistics of thecalibration and validation results Overall the calibratedresults are improved over the original VIC simulations in L13The Sabine and Neches Basins where there is ample rainfalland runoff have the best calibration results among all of thebasins studied The Brazos River Basin which has the largestdrainage area does not match the observations well duringthe low flow seasons (AugustndashNovember) A possible reasonfor this is that the Brazos River is highly regulated by manyreservoirs which may have altered the streamflow patterns

significantly To test this the VIC simulated streamflowwas compared with observations during the prereservoirera and the postreservoir era Because most reservoirs onthe Brazos were built after the 1960s results from 1939 to1960 were considered prereservoir (as observed streamflowrecord started in 1939) and results from 1961 to 2011 wereconsidered postreservoir It was found that the 1198772 and NSEvalues are 088 and 064 prereservoir while the values are 084and 062 postreservoir Given that VIC simulated flows arenaturalized flows (ie no reservoir effects are considered)such discrepancy before and after reservoir construction isunavoidable Regardless the error statistics for the Brazoshave improved the most (of all the basins in the study)Although the calibrated streamflow over the Nueces doesnot outperform the L13 results (in terms of all four ofthe statistical variables) its annual cycle and MAE haveshown much better agreement with the observations than


Table 2 The statistics of calibrated and validated monthly flows

Basin Conditions Period 1198772 NSE MAE119874119861119878 RMSE119874119861119878

SABINL13 1960ndash1985 087 069 027 055

Calibration 1960ndash1985 088 076 003 048Validation 1925ndash2011 088 076 005 050

NECHEL13 1960ndash1985 081 057 018 065


TRNTYL13 1960ndash1985 083 068 005 060


BRAZOL13 1960ndash1985 062 023 035 099


COLORL13 1960ndash1985 061 046 076 120


GUADAL13 1960ndash1985 077 052 026 071


SANANL13 1960ndash1985 083 059 034 085


NUECEL13 1960ndash1985 086 062 085 166


SANJAL13 1960ndash1985 075 053 008 095


LAVACL13 1960ndash1985 082 054 022 111


the L13 dataset does Indeed the calibration has successfullyeliminated the overestimation in the September and October(shown by the L13) dataset over the Nueces Basin

25 Model Validation The performance of the VIC simula-tions was evaluated in terms of streamflow and soil moistureresults The former is the most commonly adopted approachfor testing water budget terms as a whole The latter is ofspecial importance since soil moisture was used to quantifydroughts in this study Such comprehensive comparisonsallow us to sufficiently test the robustness of this dataset

Firstly the streamflow values simulated using the opti-mally calibrated parameter sets were validated over eachbasin based on the availability of USGS streamflow observa-tions Overall the validation results (in Table 2) are consistentwith the calibration across all basins The 1198772 and NSE valuesfor the calibration period range from 077sim091 and 050sim078 while the 1198772 and NSE for the validation period rangefrom 072sim088 and 045sim076 The best performance (with

regard to validation) is found at the Sabine and Neches Riverbasins while the worst is at the Nueces River Basin

Secondly the modeled soil moisture was compared within situ observations The quality controlled observationalsoil moisture data from the North American Soil MoistureDatabase (NASMD) [62] was adopted for validating the VICsimulated soil moisture Currently NASMD includes datafrom 27 observational networks and 1800 sites across NorthAmerica Here NASMD soil moisture observations from 31sites located in Texas (Figure 1) were used to evaluate the VICmodel simulated soil moisture products In this study soilmoisturewas simulated at 18∘ resolution over three soil layersoccurring at depths of 0ndash10 cm 10ndash40 cm and 40ndash100 cmrespectively The NASMD in situ observations were collectedat 5 cm and 25 cm depths The VIC soil moisture outputsat the top layer were validated by the top layer NASMDin situ observations and the VIC outputs at the middlelayer were compared with the observations made at 25 cmConsidering the different scales of the point observations andthe gridded simulations the averaged soil moisture values


Table 3 Validation results for the simulated soil moisture

Error metrics(daily 2003ndash2010)

OBS 5 cm (top layer) OBS 25 cm (second layer)SIMlowast L13 SIMlowast L13

1198772 075 073 075 071

RMSE (m3mminus3) 00349 00421 00206 00285Bias (m3mminus3) 00313 00395 minus00146 minus00185Bias119877() 1670 2113 minus642 minus811

lowastSimulated results from this study

257 260 263 26626

28

31

33N

E

PREC trend JJA

minus0024

minus0016

minus0008

0000

0008

0016

0024

257 260 263 26626

28

31

33

(mm

mon

th)

(mm

mon

th)

N

E

PREC trend DJF

minus0024

minus0016

minus0008

0000

0008

0016

0024

(a)

257 260 263 26626

28

31

33

N

E

TMAX trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMAX trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(b)

257 260 263 26626

28

31

33N

E

TMIN trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMIN trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(c)

Figure 3 Summer (JunendashAugust) andwinter (DecemberndashFebruary) precipitation (a)maximum temperature (b) andminimum temperature(c) trend

from the 31 reporting NASMD sites were compared withthe averaged VIC soil moisture values from the 31 gridsoverlaying those sites This spatial averaging approach hasbeen commonly adopted for evaluating a remotely sensed (ormodeled) soil moisture product using in situ observations[62 63]

Statistical metricsmdashincluding the Root Mean SquaredError (RMSE) the Bias and the Bias ratiomdashwere usedto determine the errors associated with the simulated soilmoisture Table 3 suggests that the soil moisture errormetricshave been improved at both layers when compared with theL13 dataset

3 Results and Applications

In this section the VIC simulated hydrologic records areused in three applications (1) investigating the changes inthe climate and hydrologic cycles between two historicalperiods (2) characterizing historical drought events usingreconstructed soil moisture information and (3) exploring

the capability of quantifying both peak flows and the recur-rence intervals of flood events from simulated peak flows

31 Changes of the Hydrologic Cycle Over the entire domainwe first examined the trends of the gridded meteorologicalforcings for summer and winter (Figure 3) Summer (June-July-August JJA) precipitation decreased across the entirestate of Texas with the exception of the northwest corner Incontrast winter (December-January-February DJF) precip-itation increased in the semiarid mid-Texas and west Texasregions but decreased in the humid east Texas region Themaximum temperature increased in most of Texas duringboth seasonsmdashwith summer being the largest in magnitudeThe minimum temperature also increased in both summerand winter Compared to the maximum temperature trendthe changes with minimum temperature are relatively small(but are more uniform)

The annual cycles of the water budget terms over thetwo historical periods were then compared over each basin(Figure 4) Most Texas river basins are characterized by


R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)

J F M A M J J A S O N D0

20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)

J F M A M J J A S O N D

NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC

Figure 4 Annual cycle of surface hydrology (P = precipitation E = evapotranspiration and R = runoff + base flow)

two precipitation peaks (one in the spring and one in thefall) with very little rainfall during the summer FromPeriod 1 to Period 2 precipitation has increased across allof the basins studied with the largest changes occurringduring the peak months Among these basins a notableincrease of precipitation is captured in the San Jacinto andLavaca basins during Period 2 The Brazos and ColoradoRiver Basins which are the two largest basins have lessprecipitation and much smaller runoff than the other basinsEvapotranspiration has only one peak which occurs in Maydue to the coinciding high soil moisture and the warmtemperature With regard to runoff the smallest values arefound in August and SeptemberThe Sabine and Neches bothgenerate more winter runoff than the other basins Drivenby precipitation changes runoff also increases during Period2 As explained earlier about the impact of the temperaturetrend the warming in Period 2 has little effect on alteringrunoff Texas is thus prone to both droughts and floods asa consequence of the large seasonal variations in the waterbudget terms

32 Drought Analysis From 1918 to 2011 there were fiveremarkably severe droughts in Texas The 1925 drought setrecord high temperatures and record low rainfall From 1930to 1936 the famous Dust Bowl drought led to tremendouseconomic and agricultural losses The catastrophic 1950sdrought lasted for seven years (1950ndash1957) and subsequentlyhas been considered the worst drought event in Texas In1971 some portions of North Texas received only one inch(254 cm) of rainfall during the entire year As a resultthis severe drought cost $100 million worth of crop losses(mainly with wheat and cotton) and killed over 100000 cattle(due to the drying up of grasslands and thirst from hightemperatures) In 2011 the region experienced the hottest

and driest one-year period ever recorded with a loss of $762billion in the agriculture sector alone [10 64 65]

In this section the hydrologic records provided by theVIC simulations are used to offer new perspectives on thesedrought events particularly focusing on agricultural droughtFigure 5 shows the drought outlook over the entire domainusing the time series values of precipitation temperature soilmoisture anomaly runoffprecipitation ratio (119877119875) droughtseverity and drought areal extent

As a function of both precipitation and temperature thePalmer Drought Severity Index (PDSI) is a very commonlyused index for detecting meteorological drought [66 67]However whether PDSI represents soil moisture conditionsis still debatable A study by Dai et al [68] concluded thatPDSI does not reflect soil moisture conditions and thereforeis not a goodmeasure of agricultural drought but others havefound that the PDSI correlates quite well with the observedand modeled monthly soil moisture contents over a largescale [69] The main advantage in using the soil moisturebased index to monitor agricultural drought is that soilmoisture deficit is affected by bothmeteorological conditions(ie precipitation and temperature) and by soilvegetationtypes Unlike PDSI this index can provide soil moistureinformation that is directly useful for water managementunder drought conditionsThe disadvantage of this approachis that accurate soil moisture data are hard to acquire Onthe one hand in situ measurements are spatially and tempo-rally limited making it challenging for monitoring droughtconsistently at a large scale On the other hand modeled soilmoisture datasets are typically not systematically evaluatedHowever by using the modeled soil moisture which hasbeen validated by in situ measurements these limitations areovercome in this study

In this study an agricultural drought is defined usingthe 10th percentile of monthly soil moisture in a grid cell


1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins

Monthly total precipitation anomaly (m

mm

onth

)Pr

ecip

itatio

n an

omal

y

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus4minus3minus2minus1

01234 Monthly mean soil moisture anomaly

SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008

minus20minus15minus10minus05

0005101520 Monthly mean temperature anomaly Drought severity

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008000102030405060708 Monthly runoffprecipitation ratio

RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)

Drought areal extent

Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)

Figure 5 20th century Texas drought outlook (climate surface hydrology drought severity and drought areal extent)

as a threshold [70] The drought severity is calculated as theproduct of the monthly soil moisture deficit () and theduration (counting the number of months that experiencedrought) The drought extent is calculated for each yearrepresented by the percentage of grid cells that experience atleast one month of drought Both the 1956 and 2011 severedroughts stand out clearly mainly because precipitation the119877119875 ratio and the soil moisture anomaly were all at recordlows and temperature set record highs Overall the five mostsevere droughts are well captured by the simulated droughtoutlook

Figure 6 shows the spatial patterns of drought severity andduration for the five selected historical drought events (in theorder of severity 1956 2011 1925 1934 and 1971)The severityand durationmaps tend to share a similar spatial patternThe1956 drought was the most catastrophic due to its severityand long duration The 2011 drought was the most severesingle year drought while the 1925 drought was characterizedby its long duration The region with the largest drought

severity is centered on eastern Texas in 1925 while the highestimpact drought is the one in the Trinity River basin in 1934Drought is hardly detected in the Upper Colorado basin andin southern Texas during 1934 The drought in 1971 was theleast severe among these five events with the area affectedlocated in the San Antonio and lower Colorado River basinsThemaximumdrought durations are associatedwith the 1956and 1925 droughts According to the analysis of the five severedrought events the Colorado River basin and the regionalong the Gulf coast are more vulnerable to drought than theother areas

33 Flood Analysis An annual maximum series analysis(AMS [20]) was performed to investigate the magnitudeand recurrence interval of flood events The AMS of a givenyear is the maximum daily streamflow value that occurredin that year In this study there are 94 AMS values duringthe entire simulation period (1918ndash2011) for each basin Twosets of AMS values were calculated for the 10 basins based


257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E

Figure 6 Reconstructed drought severity and duration

minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE

Figure 7Annualmaximumstreamflow (AMS) anomaly () duringthe period from 1918 to 2011

on daily streamflow from USGS observations and from VICsimulations

Figure 7 shows the comparison of the relative AMSanomaly (in terms of percentage) between observations andmodel simulations The relative AMS anomaly is calculatedby dividing the anomaly value with the mean AMS Themean AMS for a basin of interest is the averaged value ofthose 94 AMS values We used the relative AMS anomalyto make the basins comparable because each basin has itsown range of AMS Overall the simulated AMS values arein agreement with the observed ones The median and theminimum values of the simulated AMS anomaly are largerthan the observationsmdashbut the range of the simulated AMSanomalies is smaller than its observed counterpart in mostcases The differences between the modeled and observedAMS anomalies are mainly attributed to two factors firstthe model was calibrated using criteria based on monthlystreamflow while the AMS anomalies are statistics fromdaily data Second the gridded precipitation forcings usually

underestimate the extreme values especially over regionslike Texas where the rate of rainfall can be very large overa short period of time [71 72] The San Antonio Nuecesand Lavaca river basins (where the basin size in eachcase is relatively small compared to other basins) tend tohave larger interannual variability in AMS The five riverbasins with the largest AMS anomalies are the San AntonioNueces Lavaca San Jacinto and Guadalupe These basinsare relatively small in size and they are primarily locatedalong the coast of central Texas Driven by large seasonaland interannual precipitation variations the AMS anomaliesare therefore substantial These basins are very prone tofloodsmdashincluding hurricane floods due to their vicinity tothe coast The simulated maximum AMS results best agreewith observations over the Guadalupe and San Jacinto Riverbasins

With regard to flood analysis it is essential to understandthe relationship between the magnitude of peak events andtheir frequency of occurrence (in terms of return period)Theconcept of return period 119879 is used to describe the likelihoodof occurrences [73] An extreme event is defined as occurringwhen a random variable 119883 is greater than or equal to acertain level 119909119879The recurrence interval 120590 is the time betweenoccurrences of 119883 ge 119909119879 Here we define 119909119879 as the 90thpercentile 80th percentile and 50th percentile of the annualmaximum time series which are associated with a recurrenceinterval of 10 5 and 2 years respectively According toTable 4 the simulated and observed recurrence intervalsare in good agreement especially for the shorter recurrenceintervals The simulated flows tend to be underestimated atthe 90th percentile of AMS which leads to an overestimationof the 10-year recurrence interval This is largely due to twofactorsmdashthe calibration using monthly data and the fact thatgridded forcings tend to underestimate precipitation duringfloods

Figure 8 shows the return period of all the AMS values(from 1918 to 2011) over each basin The Brazos River Basinhas the largest AMS values for all return periods This basinhas the largest drainage area and the mean value of AMS


Table 4 Peak flow recurrence interval

BasinRecurrence interval (year)

Above 90th percentile of AMS Above 80th percentile of AMS Above 50th percentile of AMSOBS SIM OBS SIM OBS SIM

SABIN 96 106 45 46 20 20NECHE 33 88 39 48 18 19TRNTY 66 84 24 24 20 19BRAZO 80 99 38 40 16 16COLOR 94 94 37 38 16 16GUADA 80 76 44 44 20 20SANAN 90 90 49 49 20 20NUECE 90 101 49 51 20 20SANJA 90 96 45 48 19 19LAVAC 99 94 46 48 20 20Average 82 93 42 44 19 19

1 10 100 100010

100

1000

SABINNECHETRNTYBRAZOCOLOR

GUADASANANNUECESANJALAVAC

Return period (yr)

Annual maximum streamflow

(m3 s

)

Figure 8 Return period of annual maximum streamflow from thesimulated streamflow

(1482m3s) is nearly two times larger than that of the SabineBasin (which has the second largest mean AMS at 684m3s)The two river basins with the smallest AMS values for a givenreturn period are the San Jacinto and the Lavaca

4 Discussion and Summary

Wehave produced amodel simulated hydrological dataset forthe period of 1918ndash2011 at 18∘ spatial resolution over 10 Texasriver basins Because all of the basins are in juxtapositionthey share similar meteorological conditions In this waywhen one basin suffers drought or flood the neighboring

basins have a good chance of experiencing similar conditionsThe basins are correlated but they are hydrologically inde-pendent Since basin boundaries are delineated according tothe Digital Elevation Model (DEM) water from one basindoes not naturally move to the neighboring basins unlessthere is water management involved (eg an interbasin watertransfer) When comparing the basinsrsquo correlations underextreme conditions neighboring basins are more likely toexperience drought at the same time than flood This isbecause droughts usually occur over a large area (due toa lack of precipitation over several months as shown inFigure 6) while floods have large spatial heterogeneity butshort durations

The simulated streamflow was for the first time to ourknowledge calibrated and validated against USGS stream-flow observations at each basin Furthermore the modeledsoil moisture results were evaluated against in situ observa-tions Even though the VIC modeled soil moisture showswetter conditions than the observed soil moisture the cor-relation coefficient and the error values have been improvedover previous studiesThese reliable andwell evaluated resultsare expected to contribute to water resources managementagricultural planning and many other related fields in Texas

In this study we explored some applications of this newdataset by analyzing changes in water budget terms andby investigating new perspectives related to hydrologicalextreme eventsThe seasonal cycles of the water budget termsare very dynamic for all of the basins which confirms thatthe region is prone to both droughts and floods Overall thesimulated droughts are in good agreement with documentedhistorical droughtsThe soilmoisture data also provide a basisfor better depicturing drought duration and many othercharacteristicsmdashquantitativelymdashin time and space

An AMS approach was used to study flooding eventsHowever because of the intrinsic complexity and short termnature of floods (which occur on a timescale of hours todays) the simulation does not perform as well as it doeswith droughts This can be partially attributed to the fact


that the model calibration was implemented at a monthlytime scale to minimize the long-term differences between theobserved and simulated streamflowThereforemodeling skillin representing daily peak discharge is limited A daily stepor an event-based calibration will likely result in an improveddataset for investigating floods (but this would need to besubstantiated via another study) Another possible limitingfactor (with regard to the use of this dataset for simulatingfloods) is that reservoir flood control activities were notconsidered in our simulations Even though this calibratedmodel has a limitation with regard to capturing extremeflood events precisely it can still provide useful informationfor assisting planning and decision making for future watermanagement activities Nevertheless given the fast growthof the state of Texas and the continuously changing climatethis well evaluated dataset may serve as a benchmark forinvestigating the evolution of hydrological processes andextreme events in the future For instance by driving thecalibrated model in this study with multiple future scenariosavailable from the Coupled Model Intercomparison ProjectPhase 5 (CMIP5)mdashwhich has projections until 2099 and thesame spatial resolution as the VICmodelmdashstreamflow undera changing climate in these basins can be projected

Competing Interests

The authors declare that they have no competing interests

Acknowledgments

This study was performed under the sponsorships of theUS National Science Foundation Grant CBET-1454297 andthe Collaborative Research Grant Program from Texas AampMUniversity and the Consejo Nacional de Ciencia y Tecnolo-gia (TAMU-CONACYT 2014-028) Kyungtae Lee is par-tially sponsored by the Mills Scholarship 2015-16 from theTexas Water Resources Institute Maoyi Huang is supportedby the Integrated Assessment Research program throughthe Integrated Multi-Sector Multi-Scale Modeling ScientificFocus Area sponsored by the Biological and EnvironmentalResearch Division Office of Science US Department ofEnergy PNNL is operated by Battelle Memorial Institute forthe US Department of Energy under Contract DE-AC05-76RLO1830 The authors thank Dr Do Hyuk Kang fromthe NASA Goddard Space Flight Center who gave themtechnical suggestions about the model The authors alsothank Dr Ben Livneh from the Cooperative Institute forResearch in Environmental Sciences (CIRES) University ofColorado who provided the long-term hydrologic datasets asa baseline

References

[1] T J Larkin and G W Bomar Climatic Atlas of Texas vol 3Texas Department of Water Resources 1983

[2] B Guerrero ldquoThe impact of agricultural drought losses on theTexas economy 2011rdquo Briefing Paper AgriLife Extension 2012

[3] C S Gleaton and C G Anderson Facts about Texas andUS Agriculture Texas Cooperative Extension Department of

Agricultural Economics The Texas AampM University SystemCollege Station Tex USA 2005

[4] D N Fernando K C Mo R Fu et al ldquoWhat caused the springintensification and winter demise of the 2011 drought overTexasrdquo Climate Dynamics pp 1ndash14 2016

[5] R M Rauber J E Walsh and D J Charlevoix Severe andHazardous Weather KendallHunt 2008

[6] S D Schubert M J Suarez P J Pegion R D Koster and JT Bacmeister ldquoCauses of long-term drought in the US greatplainsrdquo Journal of Climate vol 17 no 3 pp 485ndash503 2004

[7] R Seager Y Kushnir C Herweijer N Naik and J VelezldquoModeling of tropical forcing of persistent droughts and pluvialsover western North America 1856ndash2000rdquo Journal of Climatevol 18 no 19 pp 4065ndash4088 2005

[8] FEMA National Mitigation Strategy Partnerships for BuildingSafer Communities Mitigation Directorate Federal EmergencyManagement Agency Washington DC USA 1995

[9] D A Wilhite M D Svoboda and M J Hayes ldquoUnderstandingthe complex impacts of drought a key to enhancing droughtmitigation and preparednessrdquo Water Resources Managementvol 21 no 5 pp 763ndash774 2007

[10] J W Nielsen-Gammon ldquoThe 2011 Texas droughtrdquo Texas WaterJournal vol 3 no 1 pp 59ndash95 2012

[11] X Dong B Xi A Kennedy et al ldquoInvestigation of the 2006drought and 2007 flood extremes at the Southern Great Plainsthrough an integrative analysis of observationsrdquo Journal ofGeophysical Research Atmospheres vol 116 no 3 2011

[12] C G Collier ldquoFlash flood forecasting what are the limits ofpredictabilityrdquo Quarterly Journal of the Royal MeteorologicalSociety vol 133 no 622 pp 3ndash23 2007

[13] T Funk ldquoHeavy convective rainfall forecasting a look atelevated convection propagation and precipitation efficiencyrdquoin Proceedings of the 10th Severe Storm and Doppler RadarConference Des Moines Iowa USA March 2006

[14] M W Downton J Z B Miller and R A Pielke Jr ldquoReanalysisof US National Weather Service flood loss databaserdquo NaturalHazards Review vol 6 no 1 pp 13ndash22 2005

[15] H O Sharif T Jackson M Hossain S B Shafique and DZane ldquoMotor vehicle-related flood fatalities in Texas1959ndash2008rdquo Journal of Transportation Safety and Security vol 2 no4 pp 325ndash335 2010

[16] H O Sharif T L Jackson M M Hossain and D ZaneldquoAnalysis of flood fatalities in texasrdquo Natural Hazards Reviewvol 16 no 1 Article ID 4014016 2015

[17] C M Goodess ldquoHow is the frequency location and severityof extreme events likely to change up to 2060rdquo EnvironmentalScience amp Policy vol 27 S1 pp S4ndashS14 2012

[18] G Luber and M McGeehin ldquoClimate change and extreme heateventsrdquo American Journal of Preventive Medicine vol 35 no 5pp 429ndash435 2008

[19] K E Trenberth J T Fasullo and T G Shepherd ldquoAttributionof climate extreme eventsrdquoNature Climate Change vol 5 no 8pp 725ndash730 2015

[20] G Zhao H Gao and L Cuo ldquoEffects of urbanization andclimate change on peak flows over the San Antonio River BasinTexasrdquo Journal of Hydrometeorology vol 17 no 9 pp 2371ndash23892016

[21] R A Wurbs and R A Ayala ldquoReservoir evaporation in TexasUSArdquo Journal of Hydrology vol 510 pp 1ndash9 2014

[22] Y Xia M B Ek C D Peters-Lidard et al ldquoApplication ofUSDMstatistics inNLDAS-2 optimal blendedNLDASdrought


index over the continental United Statesrdquo Journal of GeophysicalResearch Atmospheres vol 119 no 6 pp 2947ndash2965 2014

[23] E Etienne N Devineni R Khanbilvardi andU Lall ldquoDevelop-ment of a Demand Sensitive Drought Index and its applicationfor agriculture over the conterminous United Statesrdquo Journal ofHydrology vol 534 pp 219ndash229 2016

[24] Z Hao F Hao Y Xia et al ldquoA statistical method for categoricaldrought prediction based on NLDAS-2rdquo Journal of AppliedMeteorology and Climatology vol 55 no 4 pp 1049ndash1061 2016

[25] B Livneh and M P Hoerling ldquoThe physics of drought in theUS central great plainsrdquo Journal of Climate vol 29 no 18 pp6783ndash6804 2016

[26] N S Christensen and D P Lettenmaier ldquoA multimodel ensem-ble approach to assessment of climate change impacts on thehydrology and water resources of the Colorado River BasinrdquoHydrology andEarth SystemSciences vol 11 no 4 pp 1417ndash14342007

[27] N S Christensen AWWoodN Voisin D P Lettenmaier andR N Palmer ldquoThe effects of climate change on the hydrologyand water resources of the Colorado River basinrdquo ClimaticChange vol 62 no 1ndash3 pp 337ndash363 2004

[28] E P Maurer A W Wood J C Adam D P Lettenmaier andB Nijssen ldquoA long-term hydrologically based dataset of landsurface fluxes and states for the conterminous United StatesrdquoJournal of Climate vol 15 no 22 pp 3237ndash3251 2002

[29] B Livneh E A Rosenberg C Lin et al ldquoA long-term hydro-logically based dataset of land surface fluxes and states for theconterminous United States update and extensionsrdquo Journal ofClimate vol 26 no 23 pp 9384ndash9392 2013

[30] A A Oubeidillah S-C Kao M Ashfaq B S Naz andG Tootle ldquoA large-scale high-resolution hydrological modelparameter data set for climate change impact assessment for theconterminousUSrdquoHydrology and Earth System Sciences vol 18no 1 pp 67ndash84 2014

[31] T M Kimmel J Nielsen-Gammon B Rose and H M MogilldquoTheweather and climate of texas a big state with big extremesrdquoWeatherwise vol 69 no 5 pp 25ndash33 2016

[32] S W Lyons ldquoSpatial and temporal variability of monthlyprecipitation in Texasrdquo Monthly Weather Review vol 118 no12 pp 2634ndash2648 1990

[33] G W Bomar Texas Weather University of Texas Press 1995[34] Bureau of Economic Geology River BasinMap of Texas Bureau

of Economic Geology Austin Tex USA 1996[35] USDA-NASSCensus of Agriculture USDepartment of Agricul-

ture National Agricultural Statistics Service Washington DCUSA 2007

[36] Xu Liang D P Lettenmaier E F Wood and S J BurgesldquoA simple hydrologically based model of land surface waterand energy fluxes for general circulation modelsrdquo Journal ofGeophysical Research vol 99 no 7 pp 14415ndash14428 1994

[37] H Gao Q H Tang C R Ferguson E F Wood and D PLettenmaier ldquoEstimating the water budget of major US riverbasins via remote sensingrdquo International Journal of RemoteSensing vol 31 no 14 pp 3955ndash3978 2010

[38] I Haddeland T Skaugen and D P Lettenmaier ldquoHydrologiceffects of land and water management in North America andAsia 1700ndash1992rdquo Hydrology and Earth System Sciences vol 11no 2 pp 1035ndash1045 2007

[39] B Nijssen G M OrsquoDonnell D P Lettenmaier D Lohmannand E F Wood ldquoPredicting the discharge of global riversrdquoJournal of Climate vol 14 no 15 pp 3307ndash3323 2001

[40] HWu J S Kimball MM Elsner NMantua R F Adler and JStanford ldquoProjected climate change impacts on the hydrologyand temperature of Pacific Northwest riversrdquo Water ResourcesResearch vol 48 no 11 2012

[41] F Zhao F H S Chiew L Zhang J Vaze J-M Perraudand M Li ldquoApplication of a macroscale hydrologic modelto estimate streamflow across Southeast Australiardquo Journal ofHydrometeorology vol 13 no 4 pp 1233ndash1250 2012

[42] J Chang H Zhang YWang and Y Zhu ldquoAssessing the impactof climate variability and human activities on streamflowvariationrdquo Hydrology and Earth System Sciences vol 20 no 4pp 1547ndash1560 2016

[43] X Yuan ldquoAn experimental seasonal hydrological forecastingsystem over the Yellow River basinmdashpart 2 the added valuefrom climate forecast modelsrdquo Hydrology and Earth SystemSciences vol 20 no 6 pp 2453ndash2466 2016

[44] K M Andreadis and D P Lettenmaier ldquoTrends in 20th cen-tury drought over the continental United Statesrdquo GeophysicalResearch Letters vol 33 no 10 Article ID L10403 2006

[45] J Sheffield G Goteti F Wen and E F Wood ldquoA simulated soilmoisture based drought analysis for the United Statesrdquo Journalof Geophysical Research Atmospheres vol 109 no D24 2004

[46] J Sheffield and E F Wood ldquoProjected changes in droughtoccurrence under future global warming from multi-modelmulti-scenario IPCCAR4 simulationsrdquoClimate Dynamics vol31 no 1 pp 79ndash105 2008

[47] S Shukla and A W Wood ldquoUse of a standardized runoff indexfor characterizing hydrologic droughtrdquo Geophysical ResearchLetters vol 35 no 2 7 pages 2008

[48] C Tang and T C Piechota ldquoSpatial and temporal soil moistureand drought variability in the Upper Colorado River BasinrdquoJournal of Hydrology vol 379 no 1-2 pp 122ndash135 2009

[49] R Wu and J L Kinter III ldquoAnalysis of the relationship of USdroughts with SST and soil moisture distinguishing the timescale of droughtsrdquo Journal of Climate vol 22 no 17 pp 4520ndash4538 2009

[50] L Luo J Sheffield and E Wood ldquoTowards a global droughtmonitoring and forecasting capabilityrdquo in Proceedings of the33rd NOAA Annual Climate Diagnostics and Prediction Work-shop Lincoln Neb USA October 2008

[51] J Sheffield E FWood N Chaney et al ldquoA drought monitoringand forecasting system for sub-sahara african water resourcesand food securityrdquo Bulletin of the American MeteorologicalSociety vol 95 no 6 pp 861ndash882 2014

[52] D R Cayan T Das D W Pierce T P Barnett M Tyree andA Gershunova ldquoFuture dryness in the Southwest US and thehydrology of the early 21st century droughtrdquo Proceedings of theNational Academy of Sciences of the United States of Americavol 107 no 50 pp 21271ndash21276 2010

[53] Z Guo P A Dirmeyer Z Z Hu X Gao and M ZhaoldquoEvaluation of the second global soil wetness project soilmoisture simulations 2 Sensitivity to external meteorologicalforcingrdquo Journal of Geophysical Research Atmospheres vol 111no D22 2006

[54] J SheffieldM Pan E FWood et al ldquoSnow processmodeling inthe North American Land Data Assimilation System (NLDAS)1 Evaluation of model-simulated snow cover extentrdquo Journal ofGeophysical Research D Atmospheres vol 108 no 22 2003

[55] D Lohmann R Nolte-Holube and E Raschke ldquoA large-scale horizontal routing model to be coupled to land surfaceparametrization schemesrdquo Tellus Series A Dynamic Meteorol-ogy and Oceanography vol 48 no 5 pp 708ndash721 1996


[56] D S Shepard ldquoComputer mapping the SYMAP interpolationalgorithmrdquo in Spatial Statistics and Models vol 40 of Theoryand Decision Library pp 133ndash145 Springer Dordrecht TheNetherlands 1984

[57] C Daly R P Neilson and D L Phillips ldquoA statistical-topo-graphic model for mapping climatological precipitation overmountainous terrainrdquo Journal of Applied Meteorology vol 33no 2 pp 140ndash158 1994

[58] E Kalnay M Kanamitsu R Kistler et al ldquoThe NCEPNCAR40-year reanalysis projectrdquo Bulletin of the AmericanMeteorolog-ical Society vol 77 no 3 pp 437ndash471 1996

[59] P O Yapo H V Gupta and S Sorooshian ldquoMulti-objectiveglobal optimization for hydrologic modelsrdquo Journal of Hydrol-ogy vol 204 no 1-4 pp 83ndash97 1998

[60] J E Nash and J V Sutcliffe ldquoRiver flow forecasting throughconceptual models part Imdasha discussion of principlesrdquo Journalof Hydrology vol 10 no 3 pp 282ndash290 1970

[61] E M Demaria B Nijssen and T Wagener ldquoMonte Carlosensitivity analysis of land surface parameters using theVariableInfiltration Capacity modelrdquo Journal of Geophysical ResearchAtmospheres vol 112 no 11 Article ID D11113 2007

[62] T W Ford and S M Quiring ldquoInfluence of MODIS-deriveddynamic vegetation on VIC-simulated soil moisture in okla-homardquo Journal of Hydrometeorology vol 14 no 6 pp 1910ndash19212013

[63] H Gao E F Wood T J Jackson M Drusch and R BindlishldquoUsing TRMMTMI to retrieve surface soil moisture overthe southern United States from 1998 to 2002rdquo Journal ofHydrometeorology vol 7 no 1 pp 23ndash38 2006

[64] Texas State Library and Archives CommissionMajor Droughtsin Modern Texas Texas State Library and Archives Commis-sion Austin Tex USA 2016

[65] M Waldron ldquoRains ease yearminuslong Texas droughtrdquo The NewYork Times Archives vol 59 1971

[66] W C PalmerMeteorological Drought US Department of Com-merce Weather Bureau Washington DC USA 1965

[67] M P Peters L R Iverson and S N Matthews ldquoLong-termdroughtiness and drought tolerance of eastern US forests overfive decadesrdquo Forest Ecology and Management vol 345 pp 56ndash64 2015

[68] A Dai K E Trenberth and T Qian ldquoA global dataset ofPalmer Drought Severity Index for 1870ndash2002 relationshipwith soil moisture and effects of surface warmingrdquo Journal ofHydrometeorology vol 5 no 6 pp 1117ndash1130 2004

[69] V Lakshmi T PiechotaUNarayan andC Tang ldquoSoilmoistureas an indicator of weather extremesrdquo Geophysical ResearchLetters vol 31 no 11 2004

[70] J Sheffield and E F Wood ldquoCharacteristics of global andregional drought 1950mdash2000 analysis of soil moisture datafrom off-line simulation of the terrestrial hydrologic cyclerdquoJournal of Geophysical Research Atmospheres vol 112 no 172007

[71] C-T Chen and T Knutson ldquoOn the verification and compari-son of extreme rainfall indices from climate modelsrdquo Journal ofClimate vol 21 no 7 pp 1605ndash1621 2008

[72] M Gervais L B Tremblay J R Gyakum and E AtallahldquoRepresenting extremes in a daily gridded precipitation analysisover the United States impacts of station density resolutionand gridding methodsrdquo Journal of Climate vol 27 no 14 pp5201ndash5218 2014

[73] V T ChowD RMaidment and LWMaysAppliedHydrologyMcGraw Hill 1988

Submit your manuscripts athttpswwwhindawicom

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

ClimatologyJournal of

EcologyInternational Journal of


EarthquakesJournal of


Hindawi Publishing Corporationhttpwwwhindawicom

Applied ampEnvironmentalSoil Science

Volume 2014

Mining


Journal of

Hindawi Publishing Corporation httpwwwhindawicom Volume 2014

International Journal of

Geophysics

OceanographyInternational Journal of


Journal of Computational Environmental SciencesHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal ofPetroleum Engineering


GeochemistryHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Journal of

Atmospheric SciencesInternational Journal of


OceanographyHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Advances in


MineralogyInternational Journal of


MeteorologyAdvances in

The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

Paleontology JournalHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

ScientificaHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Geological ResearchJournal of


Geology Advances in


that has reported flood related fatalities in every single yearduring that same period [16]

While battling these extreme events Texas has becomea water deficient state where the demands for fresh waterhave been exacerbated by a rapidly growing populationThesewater issues are further challenged by climatic and land usechanges both of which may alter the natural hydrologicprocesses With a changing climate hydrologic extremes areprojected to become more frequent more severe and moreuncertain [17ndash19] Additionally the increasing portion ofimpervious land cover (due to urbanization) has a directeffect on elevating flood peaks [20] Due to the importanceof water resources for Texas and its vulnerability to water-related extreme events it is necessary to understand howfuture changes may impact Texasrsquo water resources and (riversystem) water budgets [21]

In this context comprehensive and reliable hydrologicdatasets which can support the analysis of historical hydro-logic extreme events are essential Specifically high qualitydatasets can be used to identify the onset and demiseof droughts and floods along with the multiple feedbackprocesses associatedwith hydrological extremes [11] Further-more such datasets can serve as a benchmark to evaluatefuture extreme events and to prevent record setting disastersin advance (through combining effective water resourcesmanagement measures with model predictions)

With the enhanced computational capabilities a highvolume of hydrological datasets have recently been generated(and released) for studying droughts and floods For instancethe North American Land Data Assimilation System-2(NLDAS-2 [22]) includes long-term (1979ndashpresent) simu-lations of the surface hydrology for the contiguous UnitedStates at 18∘ resolution This dataset has been used forproviding long-term records of water budget terms foranalyzing historic droughts and for providing the basis forseasonal drought prediction [23ndash25] These types of datasetshave also been widely used in regional assessments of climatechange impacts on surface hydrology such as in a set ofstudies focused on the Colorado River basin [26 27] Suchmodeled hydrologic datasets have strong advantages overtraditional observation based datasets whose availability islimited in time and space For example in situ observationsare often available only at point locations or over areas muchsmaller than the model spatial resolution Comparisons aretherefore restricted to the temporal and spatial scales resolvedby the model [28]

Although many of the above-mentioned hydrologicdatasets contain gridded long-termmodeled results over theentire state of Texas the data quality is often inadequate tosupport decisionmaking Typically only a very small numberof the Texas river basins (only one or two of them) havebeen calibrated against observed streamflow (eg [28 29])The results of a study by Oubeidillah et al [30] whichcalibrated for 2107 hydrologic subbasins (8-digit hydrologicunits HUC8s) over the entire CONUS show that the Nash-Sutcliffe values for most Texas basins are negative Withouta well-tested reliable dataset all analyses will be at riskfor providing misleading conclusions and recommendationsTherefore there is a strong need for effectively constraining

the quality of hydrologic model simulation results throughcalibration over each individual river basin in Texas

Driven by the meteorological forcings of Livneh et al([29] hereafter L13) we hereby provide a calibrated andvalidated hydrological dataset for 10 major Texas river basinsThe dataset is deemed high quality because of its relativelyhigh spatial (18∘) and temporal (daily) resolutions and itsevaluated skill compared to observed hydrological variablesThe dataset includes evapotranspiration runoff and soilmoisture records from 1918 to 2011 The L13 meteorologicalforcings are available from 1915 to 2011 but we used the firstthree years for model spin-up (and then analyzed from 1918onward) The dataset generated from this study was utilizedto fulfill two research objectives (1) to evaluate the impactsof a changing climate on the water budget terms and (2)to reveal new perspectives about hydrologic extreme events(such as droughts and floods) which cannot be assessed usingtraditional observations This study is organized as followsthe data and methods are presented in Section 2 where thecalibration of the soil parameters and the validation of thesimulated results are described The quality of the simulatedsoil moisture is evaluated against observed soil moisture InSection 3 the differences in water budget terms are studiedby comparing two periods 1918ndash1959 (Period 1) and 1960ndash2011 (Period 2) Historical drought (severity and duration)and flood (recurrence interval and return period) events areinvestigated based on simulated hydrologic variablesWe alsocompare the annual cycle of the water budget at each riverbasin between the two historical periods Finally discussionand conclusions are presented in Section 4

2 Data and Methodology

21 Study Area This study focuses on the Texas Gulf Regionlocated in southern central North America (25ndash34∘N 93ndash103∘W) which has a total area of 343100 km2 The regionincludes 10 major river basins (Figure 1) and covers fiveclimate zones (from arid to subtropical humid) The domaincontains geographical properties varying from dessert (farwest Texas) to mountainous (Guadalupe Range) regions [31]The diverse climate in Texas manifests itself as large spatialand temporal variations in precipitation and temperatureThe annual mean precipitation in Southeast Texas is morethan 1400mm while Northwest Texas only receives about400mm [32] The annual mean temperature varies greatlywith latitude from north to south According to Bomar [33]the average annual temperature (1961ndash1990) in the northernportion of the Texas High Plains is 132∘C while it is 233∘Cin Southern TexasWhile most west Texas rivers flow for onlypart of the year (due to a lack of precipitation) East Texasrivers flow year-round benefiting from a subtropical climate[34]

With an annual economic revenue of $100 billion agri-culture is very important in Texas A total of 528000 km2 isoccupied by farms and ranches About 76 of Texas surfacearea is occupied by farms and ranches and 22 of this is cropland For the crop land portion about 57 is harvested 10 isgrassland and 33 is either not harvested or fails to produce



0 125 250 500 750 1000(Kilometers)













200

400

600


200

400

600


200

400

600


200

400

600


200

400

600


50

100

150

J FMAM J J A S OND0

50

100

150


50

100

150


20

40

60


20

40

60

SABIN



OBSL13SIM



(m3 s

)(m

3 s)








SABINL13 1960ndash1985 087 069 027 055


NECHEL13 1960ndash1985 081 057 018 065


TRNTYL13 1960ndash1985 083 068 005 060


BRAZOL13 1960ndash1985 062 023 035 099


COLORL13 1960ndash1985 061 046 076 120


GUADAL13 1960ndash1985 077 052 026 071


SANANL13 1960ndash1985 083 059 034 085


NUECEL13 1960ndash1985 086 062 085 166


SANJAL13 1960ndash1985 075 053 008 095


LAVACL13 1960ndash1985 082 054 022 111











1198772 075 073 075 071



257 260 263 26626

28

31

33N

E

PREC trend JJA

minus0024

minus0016

minus0008

0000

0008

0016

0024

257 260 263 26626

28

31

33

(mm

mon

th)

(mm

mon

th)

N

E

PREC trend DJF

minus0024

minus0016

minus0008

0000

0008

0016

0024

(a)

257 260 263 26626

28

31

33

N

E

TMAX trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMAX trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(b)

257 260 263 26626

28

31

33N

E

TMIN trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMIN trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(c)










R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)


20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)


NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC









1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in



0 125 250 500 750 1000(Kilometers)













200

400

600


200

400

600


200

400

600


200

400

600


200

400

600


50

100

150

J FMAM J J A S OND0

50

100

150


50

100

150


20

40

60


20

40

60

SABIN



OBSL13SIM



(m3 s

)(m

3 s)








SABINL13 1960ndash1985 087 069 027 055


NECHEL13 1960ndash1985 081 057 018 065


TRNTYL13 1960ndash1985 083 068 005 060


BRAZOL13 1960ndash1985 062 023 035 099


COLORL13 1960ndash1985 061 046 076 120


GUADAL13 1960ndash1985 077 052 026 071


SANANL13 1960ndash1985 083 059 034 085


NUECEL13 1960ndash1985 086 062 085 166


SANJAL13 1960ndash1985 075 053 008 095


LAVACL13 1960ndash1985 082 054 022 111











1198772 075 073 075 071



257 260 263 26626

28

31

33N

E

PREC trend JJA

minus0024

minus0016

minus0008

0000

0008

0016

0024

257 260 263 26626

28

31

33

(mm

mon

th)

(mm

mon

th)

N

E

PREC trend DJF

minus0024

minus0016

minus0008

0000

0008

0016

0024

(a)

257 260 263 26626

28

31

33

N

E

TMAX trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMAX trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(b)

257 260 263 26626

28

31

33N

E

TMIN trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMIN trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(c)










R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)


20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)


NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC









1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in





200

400

600


200

400

600


200

400

600


200

400

600


200

400

600


50

100

150

J FMAM J J A S OND0

50

100

150


50

100

150


20

40

60


20

40

60

SABIN



OBSL13SIM



(m3 s

)(m

3 s)








SABINL13 1960ndash1985 087 069 027 055


NECHEL13 1960ndash1985 081 057 018 065


TRNTYL13 1960ndash1985 083 068 005 060


BRAZOL13 1960ndash1985 062 023 035 099


COLORL13 1960ndash1985 061 046 076 120


GUADAL13 1960ndash1985 077 052 026 071


SANANL13 1960ndash1985 083 059 034 085


NUECEL13 1960ndash1985 086 062 085 166


SANJAL13 1960ndash1985 075 053 008 095


LAVACL13 1960ndash1985 082 054 022 111











1198772 075 073 075 071



257 260 263 26626

28

31

33N

E

PREC trend JJA

minus0024

minus0016

minus0008

0000

0008

0016

0024

257 260 263 26626

28

31

33

(mm

mon

th)

(mm

mon

th)

N

E

PREC trend DJF

minus0024

minus0016

minus0008

0000

0008

0016

0024

(a)

257 260 263 26626

28

31

33

N

E

TMAX trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMAX trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(b)

257 260 263 26626

28

31

33N

E

TMIN trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMIN trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(c)










R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)


20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)


NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC









1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in




SABINL13 1960ndash1985 087 069 027 055


NECHEL13 1960ndash1985 081 057 018 065


TRNTYL13 1960ndash1985 083 068 005 060


BRAZOL13 1960ndash1985 062 023 035 099


COLORL13 1960ndash1985 061 046 076 120


GUADAL13 1960ndash1985 077 052 026 071


SANANL13 1960ndash1985 083 059 034 085


NUECEL13 1960ndash1985 086 062 085 166


SANJAL13 1960ndash1985 075 053 008 095


LAVACL13 1960ndash1985 082 054 022 111











1198772 075 073 075 071



257 260 263 26626

28

31

33N

E

PREC trend JJA

minus0024

minus0016

minus0008

0000

0008

0016

0024

257 260 263 26626

28

31

33

(mm

mon

th)

(mm

mon

th)

N

E

PREC trend DJF

minus0024

minus0016

minus0008

0000

0008

0016

0024

(a)

257 260 263 26626

28

31

33

N

E

TMAX trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMAX trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(b)

257 260 263 26626

28

31

33N

E

TMIN trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMIN trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(c)










R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)


20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)


NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC









1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in





1198772 075 073 075 071



257 260 263 26626

28

31

33N

E

PREC trend JJA

minus0024

minus0016

minus0008

0000

0008

0016

0024

257 260 263 26626

28

31

33

(mm

mon

th)

(mm

mon

th)

N

E

PREC trend DJF

minus0024

minus0016

minus0008

0000

0008

0016

0024

(a)

257 260 263 26626

28

31

33

N

E

TMAX trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMAX trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(b)

257 260 263 26626

28

31

33N

E

TMIN trend JJA

minus0008

0006

0020

0034

0048

0062

0076

257 260 263 26626

28

31

33N

E

TMIN trend DJF

minus0008

0006

0020

0034

0048

0062

0076

(∘C

year

)(∘

Cye

ar)

(c)










R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)


20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)


NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC









1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in


R (1918~1959)R (1960~2011)

E (1918~1959)E (1960~2011)

P (1918~1959)P (1960~2011)


20406080

100120140 SABIN

(mm

mon

th)

(mm

mon

th)


NECHE


TRNTY


BRAZO


COLOR


20406080

100120140 GUADA


SANAN


NUECE


SANJA


LAVAC









1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in


1918 1928 1938 1948 1958 1968 1978 1988 1998 20080123456789

1918 1928 1938 1948 1958 1968 1978 1988 1998 20080

20

40

60

80

100

1918 1928 1938 1948 1958 1968 1978 1988 1998 2008minus60

minus40

minus20

0

20

40

60

Year

Year Year

YearYear

Year

Mean10 basins


mm

onth

)Pr

ecip

itatio

n an

omal

y



SM an

omal

y (

)1918 1928 1938 1948 1958 1968 1978 1988 1998 2008




RP

ratio

(mm

mon

th)

Dro

ught

exte

nt (

)


Tem

pera

ture

anom

aly

(∘C)

Dro

ught

seve

rity

(lowast

mon

th)







257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in


257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

seve

rity

N

E

257 260 263 26626

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925 E

N

E 257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971

N

E

SM d

efici

t (

)

000102030405060708

257 260 263 26626

28

31

33

1955ndash1957

Dro

ught

dur

atio

n

N

E257 260 263 266

26

28

31

33

2010-2011

N

E

257 260 263 26626

28

31

33

1921ndash1925

N

E

257 260 263 26626

28

31

33

1933ndash1935

N

E

257 260 263 26626

28

31

33

1969ndash1971 E

Mon

th

0

4

8

12

16

20

24N

E


minus200

minus100

0

100

200

300

400

OBSSIM

AM

S an

omal

y (

)

SABI

N

LAVA

C

SAN

JA

NU

ECE

SAN

AN

GUA

DA

COLO

R

BRA

ZO

TRN

TY

NEC

HE












1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in






1 10 100 100010

100

1000



Return period (yr)


(m3 s

)











Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in



Competing Interests


Acknowledgments


References























































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in
































































Volume 2014

Mining


Journal of



Geophysics







Journal of




Advances in











Geology Advances in

development and application of improved long...

Documents