Summer Monsoon Season Streamflow Variations in the Middle Yellow River Since 1570 CE Inferred from Tree Rings of Pinus Tabulaefo

atmosphere

Article Summer Monsoon Season Streamﬂow Variations in the Middle Yellow River since 1570 CE Inferred from Tree Rings of Pinus tabulaeformis

Feng Chen 1,2,*, Magdalena Opała-Owczarek 3 , Piotr Owczarek 4 and Youping Chen 1

1 Yunnan Key Laboratory of International Rivers and Transboundary Eco-Security, Institute of International Rivers and Eco-Security, Yunnan University, Kunming 650500, China; [email protected] 2 Key Laboratory of Tree-ring Physical and Chemical Research of China Meteorological Administration/Xinjiang Laboratory of Tree-Ring Ecology, Institute of Desert Meteorology, China Meteorological Administration, Urumqi 830002, China 3 Institute of Earth Sciences, Faculty of Natural Sciences, University of Silesia in Katowice, Ul. Bedzinska 60, 41-200 Sosnowiec, Poland; [email protected] 4 Institute of Geography and Regional Development, University of Wroclaw, Pl.Uniwersytecki 1, 50-137 Wrocław, Poland; [email protected] * Correspondence: [email protected]

Received: 25 April 2020; Accepted: 3 July 2020; Published: 6 July 2020

Abstract: This study investigates the potential reconstruction of summer monsoon season streamflow variations in the middle reaches of the Yellow River from tree rings in the Qinling Mountains. The regional chronology is significantly positively correlated with the July–October streamflow of the middle Yellow River from 1919 to 1949, and the derived reconstruction explains 36.4% of the actual streamflow variance during this period. High streamflows occurred during 1644–1757, 1795–1806, 1818–1833, 1882–1900, 1909–1920 and 1933–1963. Low streamflows occurred during 1570–1643, 1758–1794, 1807–1817, 1834–1868, 1921–1932 and 1964–2012. High and low streamflow intervals also correspond well to the East Asian summer monsoon (EASM) intensity. Some negative correlations of our streamflow reconstruction with Indo-Pacific sea surface temperature (SST) also suggest the linkage of regional streamflow changes to the Asian summer monsoon circulation. Although climate change has some important effects on the variation in streamflow, anthropogenic activities are the primary factors mediating the flow cessation of the Yellow River, based on streamflow reconstruction.

Keywords: Yellow River; tree rings; streamﬂow reconstruction; East Asian summer monsoon; climate warming

1. Introduction Hydrological changes and their associated water resources are closely related to society in China [1,2]. Hydrology-related natural disasters, such as floods and prolonged drought, can translate into infrastructure destruction and the loss of industrial and agricultural production, and have serious economic and social impacts. These impacts could seriously threaten the lifestyles of human beings in river basins in China. With the rapid increase in water consumption in China over the past 30 years and the high uncertainty of future water resource change in the context of global warming [3–6], reliable knowledge of streamflow variations is of importance for us to assess the spatial-temporal variation in water resources in river basins and remote areas. Long-term and high resolution dendrohydrological reconstructions provide important contributions to the body of knowledge on past hydrological variations and can facilitate efforts to simulate future hydrological scenarios [7–12].

Atmosphere 2020, 11, 717; doi:10.3390/atmos11070717 www.mdpi.com/journal/atmosphere Atmosphere 2020, 11, 717 2 of 11 Atmosphere 2020, 11, x FOR PEER REVIEW 2 of 11

TheThe Yellow River, River, one one of of the the longest longest river river systems systems in the in world the world and the and second the second longest longest in China, in China,provides provides abundant abundant freshwater freshwater for the forpeople the peopleliving along living the along river. the In river. the InYellow the Yellow River Riverbasin, basin,some someannually annually resolved resolved series series (12 months) (12 months) of natural of natural streamflow streamflow were were reconstructed reconstructed during during the the lastlast millennium,millennium, based on tree-ringtree-ring data fromfrom thethe TibetanTibetan PlateauPlateau [[13,14].13,14]. MostMost ofof thesethese reconstructionsreconstructions areare basedbased onon thethe responseresponse ofof tree-ringtree-ring widthwidth toto thethe streamflowstreamflow ofof waterwater years.years. SummerSummer monsoonmonsoon seasonseason streamflowstreamflow records in the middle Yellow River,River, however,however, rarelyrarely extendextend beyondbeyond thethe LittleLittle IceIce AgeAge andand contain contain relatively relatively weak weak hydrological hydrological signals signals [15–17 ].[15–17]. Some studiesSome havestudies proven have the proven sensitivity the ofsensitivity tree rings of of tree Chinese rings pineof Chinese (Pinus tabulaeformispine (Pinus )tabulaeformis to summer monsoon) to summer season monsoon streamflow season in streamflow the middle Yellowin the Rivermiddle basin Yellow [18–21 River]. The basin dendrohydrological [18–21]. The dendrohydrological potential of Chinese pine,potential however, of Chinese depends pine, on a convincinghowever, depends separation on of naturallya convincing and anthropogenically separation of naturally induced streamflowand anthropogenically signals at different induced time scalesstreamflow [14]. signals at different time scales [14]. ThisThis researchresearch focused focused on on developing developing a regional a regional chronology chronology and extractingand extracting streamflow streamflow signals signals based onbased the on tree the rings tree of ringsP. tabulaeformis of P. tabulaeformisat three at sampling three sampling sites (SR, sites XF and(SR, NWT)XF and [ 20NWT),22,23 [20,22,23]] in the Qinling in the MountainsQinling Mountains (Figure1 (Figure and Table 1 and1) andTable presented 1) and presented a streamflow a streamflow reconstruction reconstruction that overcomes that overcomes some ofsome these of restrictions, these restrictions, linked withlinked streamflow with streamflow sensitivity sensitivity and anthropogenic and anthropogenic influence, toinfluence, reconstruct to summerreconstruct monsoon summer season monsoon streamflow season variations streamflow in thevariations middle in Yellow the middle River sinceYellow 1570 River CE. since Therefore, 1570 weCE. compiled Therefore, a regionalwe compiled tree-ring a regional width tree-ring chronology width based chronology on tree-ring based data on from tree-ring three data sampling from three sites insampling the middle sites reaches in the middle of the Yellow reaches River, of the Central Yellow China. River, We Central attempted China. to We show attempted high to low to show frequency high streamflowto low frequency signals streamflow of the new signals streamflow of the reconstruction new streamflow and reconstruction discuss the effects and of discuss synoptic the variations effects of onsynoptic streamflow variations according on streamflow to the new according streamflow to the reconstructionnew streamflow and reconstruction other findings and available other findings from adjacentavailable regions. from adjacent regions.

Figure 1. Map showing the locations of sampling sites and hydrological stations in the middle reaches Figure 1. Map showing the locations of sampling sites and hydrological stations in the middle reaches of the Yellow River. of the Yellow River. Table 1. Site information for tree-ring width chronologies and hydrological stations in the middle YellowTable 1. River Site basin.information for tree-ring width chronologies and hydrological stations in the middle Yellow River basin. Number of Trees Site Code Latitude (N) Longitude (E) Elevation (m) Species Time Span (CE) Number/Sample Depthof Latitude Longitude Elevation Time Span Site Code Trees Species XF 34(N)◦29’ 110(E)◦05’ 25/51(m) 2030–2050 P. tabulaeformis 1543–2012(CE) NWT 34◦00’ 108◦58’/Sample 25 Depth/51 1380–1450 P. tabulaeformis 1795–2010 XFSR 34°29' 33◦44’ 110°05' 112◦14’ 25/51 29/57 2030–2050 1675 P. tabulaeformisP. tabulaeformis 1774–20071543–2012 HuayuankouNWT 34°00'34◦550 108°58'113◦390 25/51- 1380–1450 92 P. tabulaeformis - 1919–19881795–2010 HeishiguanSR 33°44' 34◦430 112°14'112◦560 29/57- 1675 109P. tabulaeformis - 1950–19861774–2007 Huayuankou 34°55′ 113°39′ - 92 - 1919–1988 Heishiguan 34°43′ 112°56′ - 109 - 1950–1986

Atmosphere 2020, 11, 717 3 of 11 Atmosphere 2020, 11, x FOR PEER REVIEW 3 of 11

2.2. Material Material and and Methods Methods

2.1.2.1. Geographical Geographical Settings Settings and and Tree-ring Tree-Ring Material Material TheThe study study areas areas are are situated situated in in Central Central China China with with typical typical monsoon monsoon climate climate characteristics characteristics in ina humida humid and and semi-humid semi-humid climatic climatic zone zone (Figure (Figure 1). The1). Theannual annual total totalprecipitation precipitation is approximately is approximately 633.5 mm,633.5 and mm, the and annual the average annual averagetemperature temperature is 12.5 °C, is based 12.5 on◦C, monthly based on climate monthly datasets climate of the datasets Climatic of Researchthe Climatic Unit Research(CRU) of Unitthe University (CRU) of of the East University Anglia (averaged of East Anglia over 33.5–36° (averaged N, 109–114° over 33.5–36 E, [24]).◦ N, Under109–114 the◦ E, influence [24]). Under of the the Asian influence monsoon, of the the Asian summer monsoon, climate the summerof the study climate region of the is hot study and region rainy, is whilehot and the rainy, winter while climate the winter is dry climate and cold is dry and and controlled cold and by controlled dry-cold by airflows dry-cold from airflows high fromlatitudes high (Figurelatitudes 2a). (Figure The 2meana). The annual mean streamflow annual streamflow of the Yellow of the YellowRiver (34°55 River′ (34N, ◦113°39550 N,′ 113E, 92◦39 m0 E, a.s.l.) 92 m at a.s.l.) the 3 Huayuankouat the Huayuankou hydrological hydrological station station is 1436 is m 14363/s during m /s during the period the period 1919–198 1919–1988.8. Figure Figure 2b indicates2b indicates the monthlythe monthly average average streamflow streamflow of the of themiddle middle Yellow Yellow River River from from 1919 1919 to to 1988, 1988, and and the the seasonal seasonal distributionsdistributions of of streamflow streamflow and and precipitation precipitation are are highly similar. The The stre streamflowsamflows increase increase rapidly rapidly duringduring summer, summer, along with increases increases in monsoon precipitation. The The Qinling Qinling Mountains Mountains are are the the most most importantimportant orographic orographic barrier barrier in in the the study study area area an andd serve serve as as the the dividing dividing line line between between northern northern and and southernsouthern China. China. As As a a dominant dominant tree tree species species in in the the Qinling Qinling Mountains, Mountains, Chinese Chinese pines pines grow grow on on thin thin or or rockyrocky soils soils on on the the peaks peaks of of mountains mountains and and are are known known to to be be strongly strongly sensitive sensitive to to climatic climatic change change and and areare significantly significantly positively correlatedcorrelated withwith streamflowstreamflow and and precipitation precipitation variations variations over over a broada broad area area in inCentral Central China China [19 [19–23].–23]. This This research research attempted attempted to to develop develop aa regionalregional chronologychronology and to to reconstruct reconstruct thethe streamflow streamflow of of the the Yellow Yellow River River based based on on the the tree tree rings rings of of Chinese Chinese pines pines from from three three sampling sampling sites sites (SR,(SR, XF XF and and NWT) NWT) [20,22,23] [20,22,23] located located in in the the Qinling Mountains (Figure 1 1 and and TableTable1 ).1).

Figure 2. (a) Average monthly precipitation and temperature of the middle reaches of the Yellow River Figure 2. (a) Average monthly precipitation and temperature of the middle reaches of the Yellow from 1950 to 2018, based on monthly climate datasets of the Climatic Research Unit (CRU) of the River from 1950 to 2018, based on monthly climate datasets of the Climatic Research Unit (CRU) of University of East Anglia (averaged over 33.5–36◦ N, 109–114◦ E). (b) The average monthly streamflow the University of East Anglia (averaged over 33.5–36° N, 109–114° E). (b) The average monthly from 1919 to 1988 at the Huayuankou hydrological station. (c) The average monthly streamflow from streamflow from 1919 to 1988 at the Huayuankou hydrological station. (c) The average monthly 1951 to 1986 at the Heishiguan hydrological station. streamflow from 1951 to 1986 at the Heishiguan hydrological station. Significant common signals among the individual ring-width series were confirmed by cross-dating. To removeSignificant nonclimatic common trends, signals negative among exponential the individual or straight-line ring-width curve series fits werewere used. confirmed After detrending by cross- dating.with the To negative remove exponentialnonclimatic curve,trends, all negative individual exponential ring-width or sequencesstraight-line from curve the fits three were sampling used. Aftersites weredetrending merged with into the a regionalnegative chronologyexponential (RC) curv usinge, all individual a bi-weight ring-width robust mean sequences in the ARSTANfrom the threeprogram sampling [25]. The sites bi-weight were merged robust into mean a regional minimizes chronology the effect (RC) of outliers, using a extreme bi-weight values robust or biasesmean in in thethe ARSTAN tree-ring index.program The [25]. correlation The bi-weight analysis robust revealed me thatan minimizes high correlations the effect existed of outliers, between extreme the site valueschronologies, or biases and in the regionaltree-ring chronologyindex. The correlati displayedon significant analysis revealed common that signals high (Tablecorrelations2). To captureexisted betweenthe low tothe high site frequency chronologies, common and variance,the regional the chronology standard version displayed (STD) significant of the regional common chronology signals (Tablewas used 2). To in thecapture streamflow the low reconstruction. to high frequency The common reliability variance, of the regional the standard chronology version was (STD) assessed of the by regionalthe expressed chronology population was signalused (EPS)in the and streamflow interseries re correlationconstruction. (Rbar) The [26 reliability]. Both Rbar of andthe EPSregional were chronologycomputed overwas 50assessed years, by lagged the expressed by 25 years. population To minimize signal the (EPS) influence and interseries of changing correlation sample (Rbar) depth, [26].the variance Both Rbar in and the regionalEPS were chronology computed wasover stabilized 50 years, lagged using methods by 25 years. shown To minimize by Osborn the et al.influence [27]. of changing sample depth, the variance in the regional chronology was stabilized using methods shown by Osborn et al. [27].

Atmosphere 2020, 11, 717 4 of 11

Table 2. Correlations among the tree-ring width chronologies during the common period 1850–2007.

SR XF NWT SR 1.00 * XF 0.47 * 1.00 * NWT 0.25 * 0.53 * 1.00 * RC 0.71 * 0.83 * 0.79 * * indicates 99% conﬁdence level.

2.2. Streamflow Data and Analyses We selected two hydrological stations near the sampling sites in the study area: Huanyuankou station in the Yellow River (34◦550 N, 113◦390 E, 92 m a.s.l., observation period 1919–1988) (Figure2b) and Heishiguan station in the Yiluo River (34◦430 N, 112◦560 E, 109 m a.s.l., observation period 1951–1986 (Figure2c). The monthly mean streamflow of each station was calculated. Clearly, in the middle sections of the Yellow River, high summer monsoon precipitation coincides with high streamflow during July to September, and in contrast, this area is controlled by a cold dry air mass in winter (Figure2). In Pearson’s correlation analysis, the streamflow data of Huanyuankou and Heishiguan along with the regional chronology were examined from January to December. All statistical procedures were evaluated at the p < 0.05 level of significance. Based on the results of the streamflow response, a linear regression equation was used to reconstruct the streamflow variation in the middle Yellow River. Since the calibration period is relatively short, statistical tests, including the reduction of error (RE), Pearson correlation coefficient (r) and coefficient of efficiency (CE), were computed using the method of leave-one-out cross-validation [28]. To establish connections between our streamflow reconstruction and the atmospheric circulations in East Asia, we analyzed multiple correlations of our streamflow reconstruction with the EASM index [29] and sea surface temperature dataset (HadSST2, [30]) for the common period.

3. Results and Discussion

3.1. Streamflow Reconstruction To guarantee the reliability of the streamflow reconstruction, we truncated the early portion of the regional chronology prior to 1570 based on an expressed population signal of at least 0.85. The moving EPS values from 1570 to 2012 were close to 0.9, further revealing the reliability of the regional chronology. Figure3 indicates the correlations between the monthly streamflow data and the regional chronology. The tree-ring widths of P. tabulaeformis show remarkable (p < 0.05) positive correlations with the streamflow of the Yiluo River (Heishiguan) during June–August. Similarly, there were some significant positive correlations between the regional chronology and streamflow of the Yellow River (Huayuankou) during June–December. After screening the regional chronology in correlation analysis with the streamflow combinations from January to December, the strongest correlation (r = 0.64, p < 0.01) was found between the regional chronology and the June–August streamflow of the Yiluo River, and a high correlation (r = 0.61, p < 0.01) was found between the regional chronology and the July–October streamflow of the Yiluo River. Moreover, a high correlation of the regional chronology was found for the July–October streamflow of the Yellow River, with r = 0.42 (p < 0.01). The correlation between the streamflow of Heishiguan and Huayuankou reached 0.70 (p < 0.01), despite considerable anthropogenic influences (Figure4). Atmosphere 2020, 11, x FOR PEER REVIEW 5 of 11

Atmosphere 2020, 11, 717 5 of 11 AtmosphereFigure 2020 3., 11Correlation, x FOR PEER analyses REVIEW between the regional chronology and streamflow at Heishiguan (a) 5 of 11 and Huayuankou (b). The dashed lines represent the 95% confidence level.

First, a linear regression equation was established to reconstruct the streamflow history for the Yiluo River. During the common period of the regional chronology and streamflow data (1951–1986), the streamflow reconstruction accounted for 41% of the actual streamflow variance (Figure 4a). Although this reconstruction is statistically stable, we found that streamflow in July 1958 was unusual, related to the maximum flow recorded during the instrumental period. If we eliminated 1958 from the reconstruction equation, the explained variance increases from 41% to 48%. Considering that the Yiluo River is a tributary of the Yellow River and the river basin is located in the Qinling Mountains with less human disturbances, the instrumental streamflow data of this river provideFigure a reference 4. (a) Comparison to validate between our streamflow the regional chronologyreconstruction (RC) andof the observed Yellow June–August River. To a streamflow certain extent, the regionalofFigure the Yiluo chronology4. ( Rivera) Comparison (Heishiguan). can also between represent (b) Comparison the theregional change between chros innology the the July–October natural (RC) and streamflow streamflowobserved inJune–August ofthis the region Yellow after the 1950s,Riverstreamflow (Huayuankou)as indica of theted Yiluo by and the River the results July–October (Heishiguan).. streamflow (b) Comparison of the Yiluobetween River the (Heishiguan). July–October (streamflowc) Comparison of betweenthe Yellow the River regional (Huayuankou) chronology and (RC) the and July–October the observed streamflow July–October of the streamflow Yiluo River of (Heishiguan). the Yellow River (c) (Huayuankou).Comparison between the regional chronology (RC) and the observed July–October streamflow of the Yellow River (Huayuankou). First, a linear regression equation was established to reconstruct the streamflow history for the YiluoSince River. the During 1950s, the a large common number period of hydraulic of the regional engineering chronology structures and have streamflow been built data in (1951–1986),the middle thereaches streamflow of the Yellow reconstruction River, and accounted the response for 41% of tree of therings actual to streamflow streamflow has variance been considerably (Figure4a). Althoughreduced due this to reconstruction strong anthropogenic is statistically influences stable, [14] we. found Although that streamflowthe correlation in July was 1958 high was during unusual, the relatedentire period to the maximumof 1919–1988, flow the recorded correlation during between the instrumentalthe July–October period. streamfl If weow eliminated of the Yellow 1958 River from theand reconstruction the regional chronology equation, thewas explained just 0.28 (p variance < 0.10) during increases the fromperiod 41% 1950–1988, to 48%. which Considering also indicates that the that the streamflow variation in this area is seriously influenced by anthropogenic factors. The first- Yiluo River is a tributary of the Yellow River and the river basin is located in the Qinling Mountains differenced correlation from 1950 to 1988 is r = 0.07 (p > 0.01), indicating that low frequency variation with less human disturbances, the instrumental streamflow data of this river provide a reference to is relativelyFigure 4. coherent(a) Comparison between between the regionalthe regional chronology chronology and (RC) observed and observed streamflow June–August data. The validate our streamflow reconstruction of the Yellow River. To a certain extent, the regional chronology instrumentalstreamflow streamflow of the Yiluo Riverdata (Heishiguan).are 9.7% less (b ) Comparisonthan the reconstructed between the July–October streamflow streamflow data, and of the can also represent the changes in the natural streamflow in this region after the 1950s, as indicated by anthropogenicthe Yellow River influence (Huayuankou) is even stronger and the July–Octoberin the last 30 streamflowyears [14]. ofIt theis believed Yiluo River that (Heishiguan). the streamflow (c) in the results. the middleComparison Yellow between River theis influenced regional chronology by the fluvial (RC) processesand the observed in its upper July–October reaches, streamflow such as the of Tibetan the Since the 1950s, a large number of hydraulic engineering structures have been built in the middle Plateau,Yellow and River by the(Huayuankou). streamflow from the Qinling Mountains; thus, a strong correlation (r = 0.70, p < reaches of the Yellow River, and the response of tree rings to streamflow has been considerably reduced 0.01) between the streamflow of Heishiguan and Huayuankou was found, despite considerable due toSince strong the anthropogenic 1950s, a large number influences of [hydraulic14]. Although engine theering correlation structures was have high been during built the in entire the middle period ofreaches 1919–1988, of the the Yellow correlation River, between and the the response July–October of tree streamflow rings to streamflow of the Yellow has River been and considerably the regional chronologyreduced due was to juststrong 0.28 anthropogenic (p < 0.10) during influences the period [14] 1950–1988,. Although which the alsocorrelation indicates was that high the streamflowduring the variationentire period in this of area1919–1988, is seriously the correlation influenced between by anthropogenic the July–October factors. Thestreamfl first-diowff oferenced the Yellow correlation River fromand the 1950 regional to 1988 chronology is r = 0.07 (wasp > 0.01),just 0.28 indicating (p < 0.10) that during low the frequency period 1950–1988, variation is which relatively also indicates coherent betweenthat the streamflow the regional variation chronology in andthis observedarea is seriously streamflow influenced data. The by instrumentalanthropogenic streamflow factors. The data first- are 9.7%differenced less than correlation the reconstructed from 1950 streamflow to 1988 is r data, = 0.07 and (p the> 0.01), anthropogenic indicating that influence low frequency is even stronger variation in theis relatively last 30 years coherent [14]. It isbetween believed the that regional the streamflow chronology in the middleand observed Yellow River streamflow is influenced data. by The the fluvialinstrumental processes streamflow in its upper data reaches, are 9.7% such less as thethan Tibetan the reconstructed Plateau, and bystreamflow the streamflow data, fromand the Qinlinganthropogenic Mountains; influence thus, is a strongeven stronger correlation in the (r = last0.70, 30p years< 0.01) [14]. between It is believed the streamflow that the ofstreamflow Heishiguan in andthe middle Huayuankou Yellow wasRiver found, is influenced despite by considerable the fluvial processes anthropogenic in its upper influences reaches, (Figure such4b). as the Therefore, Tibetan Plateau, and by the streamflow from the Qinling Mountains; thus, a strong correlation (r = 0.70, p < 0.01) between the streamflow of Heishiguan and Huayuankou was found, despite considerable

Atmosphere 2020, 11, 717 6 of 11 Atmosphere 2020, 11, x FOR PEER REVIEW 6 of 11 theanthropogenic tree-ring width influences chronologies (Figure of P.4b). tabulaeformis, Therefore, thewhich tree-ring are sensitive width chronologies to the streamflow of P. tabulaeformis, variation in the Qinlingwhich Mountainsare sensitive and to the were streamflow utilized asvariation a proxy in to the reconstruct Qinling Mountains the hydrological and were variables utilized as of athe proxy Yiluo River,to reconstruct may also be the used hydrological to acquire thevariables streamflow of th signalse Yiluo ofRiver, the middle may also Yellow be River.used to Detailed acquire analysis the indicatedstreamflow that signals a high of correlation the middle (r Yellow= 0.60, River.p < 0.01) Detailed was foundanalysis between indicated the that regional a high chronology correlation ( andr the= July–October0.60, p < 0.01) was streamflow found between of the Yellow the regional River chrono and afflogyected and the the changes July–October in subsequent streamflow streamflow. of the TheYellow reconstruction River and equation affected between the changes the July–October in subsequent streamflow streamflow. of the The Yellow reconstruction River and theequation regional chronologybetween the achieved July–October reliable streamflow calibration andof the cross-validation Yellow River statistics.and the regional The final chronology calibration achieved regression (Yreliable= 3462.699RC calibration775.865) and cross-validation explained 36.4% statistics. of the total The variancefinal calibration of the instrumental regression (Y streamflow = 3462.699RC during − − 1919–1949775.865) (Figureexplained4c) and36.4% 33.6% of the in total the leave-one-out variance of the cross-validation. instrumental streamflow The positive during CE (0.27) 1919–1949 and RE (0.28)(Figure revealed 4c) and the 33.6% good predictivein the leave-one-out skill of the cross-va reconstructionlidation. equation.The positive CE (0.27) and RE (0.28) revealedFigure the5a exhibitsgood predictive the interannual skill of the to reconstruction multidecadal equation. variability of the July–October streamflow reconstructionFigure 5a for exhibits the middle the interannual Yellow River to sincemultidecadal 1570 CE. variability Several extended of the July–October high and low streamflow streamflow reconstruction for the middle Yellow River since 1570 CE. Several extended high and low streamflow epochs ( 11 years) were revealed according to 25-year low pass filtered streamflow values of epochs≥ (≥11 years) were revealed according to 25-year low pass filtered streamflow values of the the reconstruction and the long-term mean (1570–2012, 2670.6 m3/s). Low streamflow epochs reconstruction and the long-term mean (1570–2012, 2670.6 m3/s). Low streamflow epochs occurred occurred around 1570–1643, 1758–1794, 1807–1817, 1834–1868, 1921–1932 and 1964–2012, while high around 1570–1643, 1758–1794, 1807–1817, 1834–1868, 1921–1932 and 1964–2012, while high streamflow periods occurred at 1644–1757, 1795–1806, 1818–1833, 1882–1900, 1909–1920 and 1933–1963. streamflow periods occurred at 1644–1757, 1795–1806, 1818–1833, 1882–1900, 1909–1920 and 1933– σ The1963. streamflows The streamflows less than less the than mean-1 the mean-1(2166.8σ (2166.8 m3/s mm) m3/s for mm) three for or three more or consecutive more consecutive years occurred years σ atoccurred 1576–1580, at 1576–1580, 1632–1635 1632–1635 and 1926–1929, and 1926–1929, and conversely, and conversely, the streamflows the streamflows higher than higher the than mean the+ 1 (3174.4mean+1 m3σ /(3174.4s) for threem3/s) orfor morethree or consecutive more consecutiv yearse occurredyears occurred at 1675–1678, at 1675–1678, 1747–1749, 1747–1749, 1803–1806, 1803– 1894–18981806, 1894–1898 and 1946–1949. and 1946–1949. A trend A trend towards towards decreased decreased streamflow streamflow begins begins in thein the middle middle of theof the 20th century20th century and continues and continues into the into 21st the century. 21st century.

Figure 5. (a) Estimated (thin line) and 25-year low pass filter (thick line) values of the July–October Figure 5. (a) Estimated (thin line) and 25-year low pass filter (thick line) values of the July–October streamflow of the Yellow River. The central horizontal line shows the mean of the estimated streamflow of the Yellow River. The central horizontal line shows the mean of the estimated values; values; the dotted lines show the border of one standard deviation. (b) Comparison between the the dotted lines show the border of one standard deviation. (b) Comparison between the July–October July–October streamflow reconstruction of the Yellow River and the EASM over the 1948–2012 period. streamflow reconstruction of the Yellow River and the EASM over the 1948–2012 period. (c) (c) Correlation patterns of the reconstructed streamflow with the July–October Indo-Pacific SSTs over the 1870–2012 period.

Atmosphere 2020, 11, 717 7 of 11

3.2. Link with the East Asian Summer Monsoon (EASM) and the Indo-Pacific Ocean Climate Commonly, with the northward movement of the monsoon rain belt, the middle Yellow River is substantially impacted by the EASM during June to October, and monsoon precipitation and streamflow increase rapidly, especially the July streamflow (Figure2). The instrumental climate data show that the total precipitation and streamflow from June to October account for 73% of the annual precipitation and 66% of the annual streamflow in the middle Yellow River basin, respectively. The correlation between the observed precipitation and streamflow from 1953 to 1988 is r = 0.60 (p < 0.01). The EASM intensity is indicated by streamflow to a certain degree. Thus, it is obvious that the streamflow variation in this region is closely linked to the intensity of the East Asian summer monsoon and its associated monsoon precipitation [31]. In other words, higher streamflows mean a stronger summer monsoon, and vice versa. Based on the above points, the low and high streamflow periods correspond to weak and strong summer monsoon intervals, respectively. To establish the linkages between our streamflow reconstruction and the EASM, we analyzed the correlations between the EASM index and our streamflow reconstruction [29] for the 1948–2012 period. The stronger correlation (r = 0.23, p = 0.06) was between the concurrent June EASM index and our streamflow reconstruction and increased to 0.56 (p < 0.01) after 5-point smoothing (Figure5b). Although we do not want to overinterpret these weak correlations, the positive correlations with the EASM index also suggest that the water cycles of the middle Yellow River are linked closely with the East Asian summer monsoon dynamics, which affect precipitation conditions over the Qinling Mountains [32–34]. This result is in line with our expectations. As shown in Figure5c, some significant negative correlation areas with SSTs were found in the northern Indian Ocean and western tropical Pacific Ocean, implying that these tropical sea areas are the main water vapor sources for the middle Yellow River. Hydroclimatic conditions in Central China would be impacted by the Asian summer monsoon through water vapor transport from sea areas around China [29,35]. These water vapor source areas influenced by the Asian summer monsoon [36,37] also suggest possible teleconnections between our streamflow reconstruction and the Asian summer monsoon systems. Significant negative correlations between the moisture-sensitive tree-ring width record and SSTs in the northern Indian Ocean and western tropical Pacific Ocean are also found in the neighboring area [36], suggesting similar hydroclimatic variations for both areas. Although instrumental climatic data showed that the EASM has significantly weakened since the 1970s [29,38], our reconstruction shows that this dry trend may have started after the warm 1940s. During the past 30 years, this downward trend has been further exacerbated by increased evaporation as a result of climate warming [39], despite the EASM recovering from its weak epoch since the 1990s [40].

3.3. Human Impacts The sampling sites used in our streamflow reconstruction in this study are relatively dry, and therefore, these pine trees have suffered from water stress during the warm period [20,22,23]. Warm season temperature increases during recent years could have contributed to increasing water scarcity in the Yellow River [39]. The temperature increase since the 1980s, shown by tree-ring records [21,41] as a period of unusual warm temperatures, could have exacerbated the recent drought across the middle Yellow River basin. At the same time, we found that there is a significant negative correlation (r = 0.31, p < 0.01) between the low frequency variations in the annual streamflow − reconstruction [14] and our seasonal streamflow reconstruction (Figure6), which may suggest that the increase in runoff caused by increased melting of snow and ice in the Tibetan Plateau under the background of climate warming cannot completely offset the decrease in runoff caused by increasing water consumption and warming-induced drought downstream. The recent drought trend, by influencing available water resources and water cycles, was also one of the contributing factors to the flow cessation events of the Yellow River that occurred in the 1990s [42,43], especially for the drought year 1997 [36]. This region is the most important grain production area in China, and as a Atmosphere 2020, 11, 717 8 of 11 result of rapid socioeconomic development, regional water consumption has rapidly increased, and the influences of human actions on the streamflow of the Yellow River are further strengthening [14]. As shown by our streamflow reconstruction, severe low runoff periods in history that coincided with drought events occurred in the middle reaches of the Yellow River, especially in the late Ming Dynasty (1620–1644), 1877–1878 and 1926–1929, causing environmental and social impacts [44–46] but rarely causing cessation of Yellow River flow. Therefore, modern flow cessation is the product of the superimposed impacts of human activities and contemporary climate change. Although no flow Atmosphere 2020, 11, x FOR PEER REVIEW 7 of 11 cessation events occurred after the implementation of the reasonable streamflow allocation scheme, a warm climate could increase evaporative demand, and the deleterious influence of climate warming Correlation patterns of the reconstructed streamflow with the July–October Indo-Pacific SSTs over will becomethe 1870–2012 considerably period. more widespread in the future [47–49]. Therefore, for the different future climate scenarios, a reasonable runoff allocation scheme is necessary.

Figure 6. Comparison between the streamflow reconstructions of the Yellow River for the summer monsoonFigure 6. seasonComparison (July–October, between thisthe streamflow study) and waterreconstr yeaructions (October–September) of the Yellow River [14 ].for the summer monsoon season (July–October, this study) and water year (October–September) [14]. 4. Conclusions 3.2. Link with the East Asian Summer Monsoon (EASM) and the Indo-Pacific Ocean Climate The regional chronology of P. tabulaeformis was constructed for the Qinling Mountain area, CentralCommonly, China. The with streamflow the northward reconstruction movement wasof th developede monsoon forrain the belt, 1570–1918 the middle period Yellow using River an equationis substantially calibrated impacted with instrumentalby the EASM streamflow during June of to the October, middle Yellowand monsoon River. Theprecipitation reconstruction and accountedstreamflow for increase 36.4% rapidly, of the especially variance inthe the July instrumental streamflow (Figure streamflow 2). The over instrumental the 1919–1949 climate period. data Theshow reconstruction that the total precipitation reflects not only and streamflow the streamflow from fluctuationsJune to October in the account middle for Yellow73% of Riverthe annual basin, butprecipitation also the strongand 66%/weak of the EASM annual variations. streamflow Significant in the middle negative Yellow correlations River basin, ofrespectively. our streamflow The reconstructioncorrelation between with the Indo-Pacific observed precipitation SSTs imply connectionsand streamflow of thefrom middle 1953 to Yellow 1988 is r River = 0.60 streamflow (p < 0.01). variationThe EASM with intensity the Asian is indicated summer by monsoon streamflow system. to a Under certain the degree. background Thus, it of is warming-drying obvious that the in recentstreamflow decades, variation streamflow in this has region been is continuously closely linked decreasing to the intensity due to recent of the intense East Asian human summer activity. Althoughmonsoon similarand its lowassociated streamflow monsoon events precipitation have occurred [31]. in In the other past, words, none of higher them causedstreamflows flow cessationmean a events.stronger Therefore, summer monsoon, reasonable and streamflow vice versa. allocation Based on isthe of above strong points, importance the low for and regional high streamflow sustainable developmentperiods correspond and the to watershed weak and environment. strong summer monsoon intervals, respectively. To establish the linkages between our streamflow reconstruction and the EASM, we analyzed the correlations Authorbetween Contributions: the EASM indexConceived and our and designedstreamflow the reconstruction experiments: F.C. [29] and for Y.C. the Performed 1948–2012 the period. experiments: The F.C.stronger and Y.C. correlation Analyzed the(r = data: 0.23, F.C. p and= 0.06) Y.C. was Contributed between reagents the concur/materialsrent/analysis June EASM tools: F.C.,index Y.C., and M.O.-O., our and P.O. Contributed to the writing of the manuscript: F.C., Y.C., M.O.-O., and P.O. All authors have read and agreedstreamflow to the publishedreconstruction version and of the increased manuscript. to 0.56 (p < 0.01) after 5-point smoothing (Figure 5b). Although we do not want to overinterpret these weak correlations, the positive correlations with the Funding: This work was supported by NSFC Project (U1803341), the National Key R&D Program of China (Grant 2018YFA0606401)EASM index also and suggest National that high-level the water talents cycles special of the support middle plan. Yellow River are linked closely with the East Asian summer monsoon dynamics, which affect precipitation conditions over the Qinling Acknowledgments: We thank the Global Runoff Data Centre for providing streamflow data. Mountains [32–34]. This result is in line with our expectations. ConflictsAs shown of Interest: in FigureThe authors 5c, some declare significant that the negative research wascorrelation conducted areas in the with absence SSTs ofwere any found commercial in the or financial relationships that could be construed as a potential conflict of interest. northern Indian Ocean and western tropical Pacific Ocean, implying that these tropical sea areas are the main water vapor sources for the middle Yellow River. Hydroclimatic conditions in Central China would be impacted by the Asian summer monsoon through water vapor transport from sea areas around China [29,35]. These water vapor source areas influenced by the Asian summer monsoon [36,37] also suggest possible teleconnections between our streamflow reconstruction and the Asian summer monsoon systems. Significant negative correlations between the moisture-sensitive tree-ring width record and SSTs in the northern Indian Ocean and western tropical Pacific Ocean are also found in the neighboring area [36], suggesting similar hydroclimatic variations for both areas. Although instrumental climatic data showed that the EASM has significantly weakened since the 1970s [29,38], our reconstruction shows that this dry trend may have started after the warm 1940s. During the past 30 years, this downward trend has been further exacerbated by increased evaporation as a result of climate warming [39], despite the EASM recovering from its weak epoch since the 1990s [40].

Atmosphere 2020, 11, 717 9 of 11

References

1. Liu, C.; Xia, J. Water problems and hydrological research in the Yellow River and the Huai and Hai River basins of China. Hydrol. Process. 2004, 18, 2197–2210. [CrossRef] 2. Du, W.; Fan, Y.; Liu, X.; Park, S.C.; Tang, X. A game-based production operation model for water resource management: An analysis of the South-to-North Water Transfer Project in China. J. Clean. Prod. 2019, 228, 1482–1493. [CrossRef] 3. Jiang, Y. China’s water scarcity. J. Environ. Manag. 2009, 90, 3185–3196. [CrossRef] 4. Jiang, S.; Wang, J.; Zhao, Y.; Shang, Y.; Gao, X.; Li, H.; Wang, Q.; Zhu, Y. Sustainability of water resources for agriculture considering grain production, trade and consumption in China from 2004 to 2013. J. Clean. Prod. 2017, 149, 1210–1218. [CrossRef] 5. Zhai, R.; Tao, F. Contributions of climate change and human activities to runoff change in seven typical catchments across China. Sci. Total Environ. 2017, 605, 219–229. [CrossRef][PubMed] 6. Liu, J.; Zhang, Q.; Singh, V.P.; Shi, P. Contribution of multiple climatic variables and human activities to streamflow changes across China. J. Hydrol. 2017, 545, 145–162. [CrossRef] 7. Gou, X.; Chen, F.; Cook, E.; Jacoby, G.; Yang, M.; Li, J. Streamflow variations of the Yellow River over the past 593 years in western China reconstructed from tree rings. Water Resour. Res. 2007, 43, 4179–4186. [CrossRef] 8. Ho, M.; Lall, U.; Sun, X.; Cook, E.R. Multiscale temporal variability and regional patterns in 555 years of conterminous U.S. streamflow. Water Resour. Res. 2017, 53, 3047–3066. [CrossRef] 9. Ravindranath, A.; Devineni, N.; Lall, U.; Cook, E.R.; Pederson, G.; Martin, J.; Woodhouse, C. Streamflow Reconstruction in the Upper Missouri River Basin Using a Novel Bayesian Network Model. Water Resour. Res. 2019, 55, 7694–7716. [CrossRef] 10. Muñoz, A.A.; González-Reyes, A.; Lara, A.; Sauchyn, D.; Christie, D.; Puchi, P.; Urrutia-Jalabert, R.; Toledo-Guerrero, I.; Aguilera-Betti, I.; Mun‘do, I.A.; et al. Streamflow variability in the Chilean

Temperate-Mediterranean climate transition (35◦ S–42◦ S) during the last 400 years inferred from tree-ring records. Clim. Dyn. 2016, 47, 4051–4066. [CrossRef] 11. Xu, C.; Buckley, B.M.; Promchote, P.; Wang, S.; Pumijumnong, N.; An, W.; Sano, M.; Nakatsuka, T.; Guo, Z. Increased Variability of Thailand’s Chao Phraya River Peak Season Flow and Its Association with ENSO Variability: Evidence from Tree Ring δ 18 O. Geophys. Res. Lett. 2019, 46, 4863–4872. [CrossRef] 12. Woodhouse, C.A.; Pederson, G.T. Investigating Runoff Efficiency in Upper Colorado River Streamflow Over Past Centuries. Water Resour. Res. 2018, 54, 286–300. [CrossRef] 13. Gou, X.; Deng, Y.; Chen, F.; Yang, M.; Fang, K.; Gao, L.; Yang, T.; Zhang, F. Tree ring based streamflow reconstruction for the Upper Yellow River over the past 1234 years. Chin. Sci. Bull. 2010, 55, 4179–4186. [CrossRef] 14. Li, J.; Xie, S.-P.; Cook, E.R.; Chen, F.; Shi, J.; Zhang, D.D.; Fang, K.; Gou, X.; Li, T.; Peng, J.; et al. Deciphering Human Contributions to Yellow River Flow Reductions and Downstream Drying Using Centuries-Long Tree Ring Records. Geophys. Res. Lett. 2019, 46, 898–905. [CrossRef] 15. Kang, L.L.; Wang, Y.Z.; Ma, Y.; Wei, Y.C. Reconstruction of the natural runoff series in recent 523 years at Huayuankou station in Yellow River. J. Water Resour. Water Eng. 2008, 19, 10–13. (In Chinese) 16. Pan, W.; Zheng, J.; Xiao, L.; Yan, F. The relationship of nature runoff changes in flood-season of middle Yellow River and Yongding River, 1766–2004. Acta Geogr. Sinica 2015, 68, 975–982. (In Chinese) 17. Pan, W.; Fei, J.; Man, Z.; Zheng, J.; Zhuang, H.; Man, Z. The fluctuation of the beginning time of flood season in North China during AD1766–1911. Quat. Int. 2015, 380, 377–381. [CrossRef] 18. Sun, J.; Liu, Y.; Wang, Y.; Bao, G.; Sun, B. Tree-ring based runoff reconstruction of the upper Fenhe River basin, North China, since 1799 AD. Quat. Int. 2013, 283, 117–124. [CrossRef] 19. Liu, N.; Bao, G.; Liu, Y.; Linderholm, H.W. Two Centuries-Long Streamflow Reconstruction Inferred from Tree Rings for the Middle Reaches of the Weihe River in Central China. Forests 2019, 10, 208. [CrossRef] 20. Shi, J.; Li, J.; Cook, E.R.; Zhang, X.; Lu, H. Growth response of Pinus tabulaeformis to climate along an elevation gradient in the eastern Qinling Mountains, central China. Clim. Res. 2012, 53, 157–167. [CrossRef] 21. Liu, Y.; Wang, Y.; Li, Q.; Song, H.; Linderholm, H.W.; Leavitt, S.W.; Wang, R.; An, Z.; Linderhlom, H.W. Tree-ring stable carbon isotope-based May–July temperature reconstruction over Nanwutai, China, for the past century and its record of 20th century warming. Quat. Sci. Rev. 2014, 93, 67–76. [CrossRef] Atmosphere 2020, 11, 717 10 of 11

22. Chen, F.; Yuan, Y.; Wei, W.; Fan, Z.; Zhang, R.; Yu, S. April–June precipitation reconstruction for Xi’an and drought assessment for the Guanzhong Plain from tree rings of Chinese pine. J. Water Clim. Chang. 2014, 6, 638–646. [CrossRef] 23. Chen, F.; Zhang, R.; Wang, H.; Qin, L.; Yuan, Y. Updated precipitation reconstruction (AD 1482–2012) for Huashan, north-central China. Theor. Appl. Climatol. 2016, 123, 723–732. [CrossRef] 24. Harris, I.; Jones, P.; Osborn, T.J.; Lister, D.H. Updated high-resolution grids of monthly climatic observations - the CRU TS3.10 Dataset. Int. J. Clim. 2013, 34, 623–642. [CrossRef] 25. Yamaguchi, D.K.; Cook, E.R.; Kairiukstis, L.A. Methods of Dendrochronology, Applications in the Environmental Sciences. Arct. Alp. Res. 1991, 23, 120. [CrossRef] 26. Fritts, H.C. Tree-Rings and Climate; Academic Press: London, UK, 1976. 27. Wigley, T.M.L.; Briffa, K.R.; Jones, P.D. On the average value of correlated time series, with applications in dendroclimatology and hydrometeorology. J. Clim. Appl. Meteorol. 1984, 23, 201–213. [CrossRef] 28. Osborn, T.J.; Briffa, K.R.; Jones, P.D. Adjusting variance for sample size in tree-ring chronologies and other regional mean time series. Dendrochronologia 1997, 15, 89–99. 29. Li, J.; Wu, Z.; Jiang, Z.; He, J. Can Global Warming Strengthen the East Asian Summer Monsoon? J. Clim. 2010, 23, 6696–6705. [CrossRef] 30. Rayner, N.A.; Horton, E.B.; Rowell, D.P.; Kent, E.; Parker, D.E.; Folland, C.K.; Alexander, L.; Kaplan, A. Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res. 2003, 108, 4407. [CrossRef] 31. Li, F.; Chen, D.; Tang, Q. Variations of hydro-meteorological variables in the Yellow River basin and their relationships with the East Asian summer monsoon. Adv. Water Sci. 2015, 26, 481–490. 32. He, H.; Zhang, Q.; Zhou, J.; Fei, J.; Xie, X. Coupling climate change with hydrological dynamic in Qinling Mountains, China. Clim. Chang. 2008, 94, 409–427. [CrossRef] 33. Tan, L.; Cai, Y.; Cheng, H.; Edwards, L.R.; Gao, Y.; Xu, H.; Zhang, Z.; An, Z. Centennial- to decadal-scale monsoon precipitation variations in the upper Hanjiang River region, China over the past 6650 years. Earth Planet. Sci. Lett. 2018, 482, 580–590. [CrossRef] 34. Chen, F.; Yuan, Y.-J.; Wei, W.-S.; Fan, Z.-A.; Yu, S.-L.; Zhang, T.; Zhang, R.; Shang, H.; Qin, L. Reconstructed precipitation for the north-central China over the past 380 years and its linkages to East Asian summer monsoon variability. Quat. Int. 2013, 283, 36–45. [CrossRef] 35. Li, H.; Dai, A.; Zhou, T.; Lu, J. Responses of East Asian summer monsoon to historical SST and atmospheric forcing during 1950–2000. Clim. Dyn. 2010, 34, 501–514. [CrossRef] 36. Fang, K.; Gou, X.; Chen, F.; D’Arrigo, R.; Li, J. Tree-ring based drought reconstruction for the Guiqing Mountain (China): Linkages to the Indian and Pacific Oceans. Int. J. Clim. 2009, 30, 1137–1145. [CrossRef] 37. Wang, L.; Qian, Y.; Zhang, Y.; Zhao, C.; Leung, L.R.; Huang, A.; Xiao, C. Observed variability of summer precipitation pattern and extreme events in East China associated with variations of the East Asian summer monsoon. Int. J. Clim. 2015, 36, 2942–2957. [CrossRef] 38. Wang, H. The weakening of the Asian monsoon circulation after the end of 1970’s. Adv. Atmos. Sci. 2001, 18, 376–386. 39. Zhai, J.; Su, B.; Krysanova, V.; Vetter, T.; Gao, C.; Jiang, T. Spatial Variation and Trends in PDSI and SPI Indices and Their Relation to Streamflow in 10 Large Regions of China. J. Clim. 2010, 23, 649–663. [CrossRef] 40. Liu, H.; Zhou, T.; Zhu, Y.; Lin, Y. The strengthening East Asia summer monsoon since the early 1990s. Chin. Sci. Bull. 2012, 57, 1553–1558. [CrossRef] 41. Chen, F.; Zhang, R.; Wang, H.; Qin, L. Recent climate warming of central China reflected by temperature-sensitive tree growth in the eastern Qinling Mountains and its linkages to the Pacific and Atlantic oceans. J. Mt. Sci. 2015, 12, 396–403. [CrossRef] 42. Ren, L.; Wang, M.; Li, C.; Zhang, W. Impacts of human activity on river runoff in the northern area of China. J. Hydrol. 2002, 261, 204–217. [CrossRef] 43. Liu, C.; Zhang, X. Causal analysis on actual water flow reduction in the mainstream of the Yellow River. Acta Geogr. Sinica 2004, 59, 323–330. (In Chinses) 44. Zheng, J.; Xiao, L.; Fang, X.; Hao, Z.; Ge, Q.; Li, B. How climate change impacted the collapse of the Ming dynasty. Clim. Chang. 2014, 127, 169–182. [CrossRef] 45. Chen, Z.; Yang, G. Analysis of drought hazards in North China: Distribution and interpretation. Nat. Hazards 2012, 65, 279–294. [CrossRef] Atmosphere 2020, 11, 717 11 of 11

46. Hao, Z.; Zheng, J.; Wu, G.; Zhang, X.; Ge, Q. 1876–1878 severe drought in North China: Facts, impacts and climatic background. Chin. Sci. Bull. 2010, 55, 3001–3007. [CrossRef] 47. Zhu, Z.; Giordano, M.; Cai, X.; Molden, D. The Yellow River Basin: Water Accounting, Water Accounts, and Current Issues. Water Int. 2004, 29, 2–10. [CrossRef] 48. Huang, J.; Zhai, J.; Jiang, T.; Wang, Y.; Li, X.; Wang, R.; Xiong, M.; Su, B.; Fischer, T. Analysis of future drought characteristics in China using the regional climate model CCLM. Clim. Dyn. 2017, 50, 507–525. [CrossRef] 49. Sternberg, T. Chinese drought, bread and the Arab Spring. Appl. Geogr. 2012, 34, 519–524. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).