Next Article in Journal
Projected Hydrological Regime Shifts in Kazakh Rivers Under CMIP6 Climate Scenarios: Integrated Modeling and Seasonal Flow Analysis
Previous Article in Journal
231Pa in the Ocean: Research Advances and Implications for Climate Change
Previous Article in Special Issue
Spatiotemporal Patterns of Hongshan Culture Settlements in Relation to Middle Holocene Climatic Fluctuation in the Horqin Dune Field, Northeast China
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Atmosphere
Volume 16
Issue 9
10.3390/atmos16091019
atmosphere-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Sedimentary Records of Paleoflood Events in the Desert Section of the Upper Yellow River Since the Late Quaternary

by
Hongli Pang
1,* and
Yunxia Jia
2
1
Key Laboratory of Ecological Safety and Sustainable Development in Arid Lands, Northwest Institute of Eco-Environment and Resources, Chinese Academy of Sciences, Lanzhou 730000, China
2
School of Geographical Science, Shanxi Normal University, Taiyuan 030031, China
*
Author to whom correspondence should be addressed.
Atmosphere 2025, 16(9), 1019; https://doi.org/10.3390/atmos16091019
Submission received: 13 July 2025 / Revised: 21 August 2025 / Accepted: 22 August 2025 / Published: 29 August 2025
(This article belongs to the Special Issue Desert Climate and Environmental Change: From Past to Present)
Browse Figures
Versions Notes

Abstract

The frequency and intensity of paleofloods reveal long-term hydrological changes and their responses to regional climate variations. This study focuses on sediment core HDZ04 from the desert section of the upper Yellow River, analyzing sediment grain size and elemental characteristics to reconstruct paleoflood events over the past 30,000 years. Using the EMMA end-member model, four end-member components were extracted, and the proportion of the two coarser end-members was used as a proxy for flood dynamics. Pearson correlation analysis indicated that ln(Zr/Ti) correlates more significantly with grain size value than ln(Zr/Rb), establishing Zr/Ti as a reliableproxy for paleoflood reconstruction. Integrating physical and chemical indicators with OSL dating, the reconstructed paleoflood sequence shows high frequency and intensity from 30~12 ka, lower values during the early and middle Holocene, and a significant increase in the late Holocene (3~0 ka). Comparison with regional climate records indicates that cold and dry periods correspond to higher paleoflood frequency and intensity. This multi-proxy approach provides a transferable framework for reconstructing past flood events in other alluvial systems worldwide, enhancing our understanding of hydrological responses to climatic forcing.

1. Introduction

Extreme flood events are essentially related to assessment of future flood risk and reliability in hydrological design. However, instrumental data are limited by the relatively short time span of observation stations, and historical documentary records are often subject to interpretive biases [1]. Paleoflood reconstruction plays an important role in extending the flooding records beyond the period covered by instrumental measurements or historical documentation [2].
The common methods in paleoflood reconstruction in the bedrock river settings involve identifying flood events, including the flood-related channel geomorphology and slackwater deposits [3,4]. Slackwater deposits are coarse-grained sediments conveyed in suspension during highly energetic flood flows and deposited in areas of flow separation after flood recession [2]. This approach has been commonly used to reconstruct extreme floods in bedrock valleys [3,5]. In non-bedrock river settings such as alluvial plains, wide valleys, and lower reaches, frequent channel migration and fine sediment accumulation complicate the identification of flood deposits due to morphological instability and altered cross-sectional profiles [6]. As such, reconstructing flood histories from fluvial sedimentary sequences often relies on the assumption that vertical variations in grain size reflect fluctuations in discharge and sediment transport capacity [7,8,9]. For instance, many researchers use various proxies (e.g., percentage of grains >306 µm, 95th percentile grain size) [5,10,11,12].
In recent years, grain-size end-member modeling (EMM) has become an increasingly important tool in paleo-environmental studies. Through quantitative unmixing of grain-size distributions into end members with distinct hydrodynamic significance, EMM provides insights into sediment transport processes and depositional environments and has been widely applied in paleoflood reconstruction [13]. However, in distal floodplain settings where vertical accretion dominates, sedimentation is often characterized by fine-grained silt and clay with limited coarse fraction input. This results in minimal grain-size contrasts between flood and non-flood deposits. Additionally, flood layers may be only a few millimeters thick, posing challenges for detection using conventional grain-size analysis with limited resolution [6]. Variations in stable elemental ratios between flood and background sediments can serve as effective paleoflood indicators [14,15,16]. The advent of high-resolution, non-destructive XRF core scanning has enabled the acquisition of continuous millimeter-scale elemental records, providing a powerful tool for reconstructing flood sequences in alluvial environments. In summary, significant progress has been made in identifying diagnostic indicators and reconstructing paleoflood events in alluvial plains.
However, the unique depositional dynamics of such systems, coupled with frequent channel migration and avulsion, often lead to sedimentary discontinuities. As a result, flood indicators may vary significantly even within the same watershed depending on site-specific geomorphic and hydrological conditions [17], making it difficult to define universally applicable flood proxies for paleoflood reconstructions across diverse catchments [18]. This underscores the need for integrated, multi-proxy approaches and comparative assessments to evaluate the reliability and applicability of various reconstruction methods in complex alluvial settings.
The Yellow River, historically known for its catastrophic floods, is one of the most intensively studied river systems globally in flood-related research. The frequency and magnitude of paleoflood events in the Yellow River exhibit pronounced temporal and spatial periodicity. In the lower reaches of the Yellow River, such events occurred predominantly during the middle to late Holocene [11,19]. Paleoflood events in the canyon section of the middle Yellow River were concentrated in the middle to late Holocene, with particularly high frequencies during two intervals, ~4.2–3.0 ka and ~3.2–3.0 ka [20,21]. The desert section, located in the upper reaches of the Yellow River, exhibits classic alluvial features with high channel mobility. Between the 6th and 18th centuries, the main stem of the Yellow River gradually shifted southward from the southern foothills of the Yin Mountains, resulting in frequent avulsions and large-scale channel migration. The surface geomorphology of the region is marked by relict oxbow lakes and a complex mosaic of ridges, plains, and depressions aligned along ancient river axes. The paleochannels are predominantly straight or slightly sinuous, trending from north-northeast to northwest [22], providing favorable geomorphic conditions for paleoflood studies. Additionally, the Hetao Plain lies at the transitional boundary between three major natural regions, making it particularly sensitive to climatic variability and an ideal region for long-term paleoflood reconstructions. In this study, we use high-resolution core sediments from the Hetao Plain to evaluate the feasibility of extracting paleoflood signals using grain-size end-member modeling and geochemical proxy analysis. This study aims to (1) reconstruct a long-term sequence of paleo-flood events in the desert section of upper Yellow River based on multi-proxy sedimentary evidence, (2) examine the intensity, frequency, and underlying driving mechanisms of paleo-floods in relation to regional climate variability, (3) situate these findings within the broader Yellow River Basin to provide implications for contemporary flood risk management and sustainable water-resource planning, and (4) offer modest insights into similar hydrological–climatic interactions in other semi-arid regions worldwide.

2. Study Area

The Hetao Plain, located in the upper reaches of the Yellow River, is a large alluvial plain situated between the Yin Mountains to the north and the Ordos Plateau to the south (Figure 1). The region exhibits a complex landscape composed of diverse landforms such as floodplains, abandoned channels, levees, shallow depressions, and relict oxbow lakes, all of which point to a history of frequent avulsion events and channel migrations. The modern Yellow River in this region exhibits a meandering to braided pattern, with strong lateral mobility and dynamic sediment transport processes. The presence of numerous paleochannels aligned in a north-northeast to northwest direction reflects the long-term wandering nature of the river over the Holocene. The Ulan Buh Desert and the Kubuq Desert are distributed along both banks of the Yellow River, forming a typical wide valley desert section of the river [23]. Climatically, the region is situated deep within the interior of the Eurasian continent and exhibits a typical temperate continental arid to semi-arid climate. Annual precipitation ranges from 150 to 363 mm, gradually increasing from west to east, and the seasonal distribution of rainfall is highly uneven [24]. Rainfall is characterized by high intensity and short duration, with over 70% of the total precipitation concentrated in the summer months, primarily influenced by the Pacific subtropical high pressure system, which triggers orographic and frontal rainfall [25].
The Hetao Plain is subject to two main types of floods: rainstorm-induced floods and ice-jam floods. In summer, short but intense convective storms can generate sudden and high-energy flood events. During winter, the river freezes, and in early spring, rapid melting of snow and ice causes abrupt increases in discharge. This process, combined with the transport of floating ice, frequently results in ice-jam floods, which pose significant hydrological and geological hazards in the region. Due to its unique geomorphic setting, dynamic depositional environment, and sensitivity to climate variability, the Hetao Plain offers a valuable natural archive for reconstructing long-term flood histories and studying fluvial responses to climatic and environmental changes in northern China. The thick sequences of fine-grained alluvial sediments deposited by the Yellow River provide ideal conditions for preserving flood-related stratigraphic records, making this region a key site for paleo-hydrological and environmental research.

3. Materials and Methods

3.1. Core Samples and X-Ray Fluorescence (XRF)

The research team conducted core drilling on the modern floodplain of the desert section of the Yellow River using a triple-tube coring system. A sediment core (HDZ04; 40°51′27.6″ N, 108°01′25″ E) was retrieved, with a total depth of 20.41 m and a recovery rate exceeding 80%. At depths from 0 to 2.36 m, the core sedimentsare characterized by a brownish-yellow clay layer with low moisture content, compact structure, and sporadic voids. At 11.73 m–12.36 m, a distinct black carbonaceous layer is present. Horizontal bedding is locally developed from 13.96 m to 20.41 m.
In the laboratory, the core was split longitudinally using a Geotek core splitter (Daventry, UK). The sediments were primarily composed of sand, silt, and clay, with little to no gravel fraction observed. Detailed sedimentological descriptions were performed during core splitting, including observations of sediment color, cohesion, sedimentary structures, and sand layer thickness. One half of the core was reserved for XRF scanning, while the other half was used for physical and chemical analyses. XRF scanning was performed using Avaatech XRF core scan at the Key Laboratory of Western China’s Environmental Systems (Ministry of Education), Lanzhou University. Details of these methods and analytical capabilities of Avaatech multifunction core scanners were described by Richter et al. [26]. Prior to scanning, the split core surface was carefully smoothed and covered with a ~4 μm thick Ultralene film. A Rhodium (Rh) tube was usedto avoid the sensitivity limitations associated with Molybdenum (Mo) and Chromium (Cr) tubes for certain elements. The scanned area was set to 6 mm (width) × 10 mm (length), with a resolution of 10 mm for continuous elemental intensity measurements [27]. Elemental compositions were measured using an XRF core scanner operated under two voltage settings. At 10 kV and 1 mA with counting time of 15 s, the instrument was capable of detecting twelve elements (Al, Si, P, S, Cl, K, Ca, Ti, Mn, Fe, Cu, and Zn). At 30 kV and 2 mA with counting time of 25 s, additional heavier elements, including Rb, Sr, and Zr, were detected [28]. Among the measured elements, Ca exhibits the highest abundance, with an average response of approximately 30,000 counts per second (cps). In comparison, the response rates of K, Ti, and Sifall within the range of ~20,000 to ~5000 cps, respectively, all of which are well above the nominal detection sensitivity of the instrument.

3.2. Grain Size Analysis

A total of 213 sediment samples were collected from drill core at 10 cm intervals. To ensure compatibility with the instrumental analytical range (0.02~2000 µm) and to prevent clogging, samples were first dry-sieved through a 2000 µm mesh to remove coarse particles, then air-dried at room temperature. For the removal of organic matter and carbonates, approximately 0.2~0.3 g of each sample was treated with 10 mL of 10% hydrogen peroxide (H2O2) and heated in a thermostatic chamber until bubbling ceased. This was followed by the addition of 10 mL of 10% dilute hydrochloric acid (HCl), with the mixture brought to a gentle boil while deionized water was added as needed to prevent evaporation. Afterward, the treated samples were topped with water and left to settle for 12 h; the supernatant was carefully siphoned off, and the lower suspension was retained. Subsequently, 10 mL of a 5% sodium hexametaphosphate ((NaPO3)6) solution was added as a dispersant, and the suspension was subjected to ultrasonic treatment for approximately 5 min. The fully dispersed samples were then analyzed for particle size distribution using the laser diffraction method.
Grain size distributions were then measured using a Malvern Mastersizer 2000 laser particle size analyzer (Malvern, UK). Sediment fractions were classified based on the Udden–Wentworth scale into clay (<3.9 μm), silt (3.9~6 μm), and sand (>6 μm). The results were divided into 68 grain size bins at 0.13Φ intervals and analyzed using the End-Member Modeling Algorithm (EMMA) to distinguish sedimentary components and infer depositional processes.

3.3. Optically Stimulated Luminescence Dating

Prior to splitting the sediment core, segments of approximately 10 cm in length were selected under darkroom conditions from relatively undisturbed fine sand layers. These samples were immediately wrapped in light-tight black plastic bags or aluminum foil to prevent light exposure. Pretreatment and measurement of the OSL samples were conducted at the Luminescence Dating Laboratory of China University of Geosciences (Wuhan). The equivalent dose (De) of each sample was determined using the Single Aliquot Regenerative-dose (SAR) protocol. For each aliquot, both natural and regenerative luminescence signals were measured, and a dose–response curve was constructed after sensitivity correction. Equivalent doses were calculated by interpolation of the natural signal onto the growth curve. The final OSL ages were computed using the software AGE•exe (2003) (Table 1).

4. Result

4.1. Grain Size Composition and Chemical Elemental Characteristics for Core Sediments

The sediment core can be divided into five units from top to bottom (Figure 2). Unit 1 is approximately 4 m thick and is primarily composed of silt, with mean grain size ranging from 6.98 μm to 43.02 μm. The sediments are characterized as poorly sorted, with the kurtosis ranges from 0.95 to 1.57, and a mean value of 1.12. Unit 2 is primarily composed of fine sand and silt, with an average particle size of 110 μm. The sediments exhibit sorting values ranging from 0.48 to 1.69 and are characterized by leptokurtic kurtosis, indicating a sharp and narrow grain size distribution. Sediments in unit 3 are finer in grain size, with silt comprising 76.08% of the total composition. The average particle size ranges from 17.97 µm to 53.54 µm, and the sediments are classified as poorly sorted. Unit 4 exhibits considerable variation in grain size and is primarily composed of fine sand and silt. The mean particle diameter ranges from 8.07 µm to 313.33 µm, with a mean of 152.1 µm. The sediments are predominantly characterized by leptokurtic distribution and are classified as poorly to very poorly sorted. Sediments in unit 5 are primarily composed of fine sand, with sorting values ranging from 0.51 to 2.61, indicating the presence of well-sorted particles.
Based on the elemental scanning results from the XRF scanner, the intensity of elements in the core sediments was obtained as a function of depth. With increasing depth, the variation trends of Ti and Rb are similar, while Zr exhibits an opposite trend. Additionally, as the grain size coarsens from unit1 to unit 2, the concentration of Zr gradually increases, whereas Ti and Rb show relatively low values.

4.2. Paleoflood Indicators Extracted from Grain Size Data

The results of the EMMA analysis show that after approximately 100 iterations, convergence was achieved at four end-members (Figure 3), indicating that four end-members represent the optimal number for this sample. From the posterior frequency distribution of the end-member components, the particle sizes of EM1 to EM3 gradually increase, with sorting improving from poor to good, while EM4 has the smallest particle size and the poorest sorting (Figure 3).For instance, the frequency curve of EM1 exhibits a bimodal distribution, with the main mode around 51 μm and a narrow and sharp shape, and the secondary mode around 328 μm. EM2 displays a unimodal distribution with a narrow and sharp curve. The particle size range is between 20–400 μm, with the mode value at 132 μm. EM3 is the coarsest of the four end-members, and its particle size frequency curve is bimodal, with the main mode at approximately 300 μm. The secondary peak shows a wide and gradual distribution pattern, with particle sizes ranging from 10–100 μm. EM4 is the finest end-member, with a bimodal frequency distribution, where the main peak is wide and flat, and the particle size range is between 1–100 μm, with a mode value of 17 μm. The secondary peak has a particle size range of 100–1000 μm, with a mode value around 190 μm. Based on modern data, sediments with a particle size greater than 80 μm in the desert section of the Yellow River are classified as bedload, while sediments with a particle size smaller than 80 μm are in suspension. Therefore, EM2 and EM3 represent the bedload components, while EM1 and EM4 correspond to suspended load components.
The values of the two coarse end-members (EM2 + EM3) reflect an increase in river flow velocity and discharge, leading to an increase in sediment particle size and content, indicating a stronger flood sediment transport capacity. As the drilling depth changes, the sum of the two coarse end-members (EM2 + EM3) shows significant fluctuations (Figure 4). Between 20.4 and 12.5 m, the sum of the end-members EM2 + EM3 shows high values, with an average of 0.97, indicating stronger hydrodynamic conditions. Between 12.5 and 10 m, the sum fluctuates significantly, with an average value dropping to 0.66. A sharp decrease is observed between 10 and 8.3 m, with the average value reaching 0.2. Between 8.3 and 4 m, the value of EM2 + EM3 rapidly increases, and then gradually decreases between 4 and 0 m, indicating weaker hydrodynamic conditions. When comparing with existing studies and the grain size indicators used in paleoflood research, the coarse end-member components show a trend that is generally consistent with changes in sand content percentage and P95. To better capture the differences in grain size components, we standardized the results of the three grain size indicators, obtaining a comprehensive grain size index curve (Figure 4), which was subsequently applied in the paleoflood event sequence reconstruction of the sediment core.

4.3. Paleoflood Proxy Indicators Extracted from Chemical Element Data

Under the influence of wind and water transport, Zr tends to be enriched in coarse-grained sediments such as zircons, which have strong resistance to weathering [29], while Ti is enriched in finer sediments such as clays [30]. Rb, on the other hand, is concentrated in fine-grained sediments that contain potassium-bearing minerals such as feldspar, mica, and clay minerals [31,32]. Therefore, the ratios of Zr/Rb and Zr/Ti reflect changes in the relative content of coarse- and fine-grained components, and these ratios are used as proxy indicators for grain size. A larger ratio indicates a higher relative content of coarse-grained components, which suggests a stronger flood magnitude [15]. However, due to differences in sedimentary environments, source materials, and transport distances between different basins, the enrichment of elements in specific grain-size components varies. As a result, the same indicator may represent different or even opposite paleo-environmental meanings in different sedimentary settings. To address this issue, Pearson correlation analysis was conducted between the elemental proxy indicators and the grain size components. Considering the asymmetry of elemental ratios, and to overcome spatial constraints and enhance differential signals [33], this study transformed the elemental ratios into the natural logarithmic ratio of elements. The results of the Pearson correlation analysis show that the coefficients of determination between ln(Zr/Ti) and the fitted grain size value is 0.4302, while the correlation coefficient between ln(Zr/Rb) and the fitted grain size value is 0.074. Therefore, compared to the ln(Zr/Rb) ratio, ln(Zr/Ti) is more suitable as a proxy for grain size and can be used to reconstruct paleoflood sequences (Figure 5).However, it should be noted that Zr/Rb may not be an ideal grain size proxy in fluvial core sediments. This is because changes in sediment source, mineral composition, or depositional environment may decouple the Zr/Rb ratio from actual grain size, introducing uncertainties when interpreting paleoflood or sedimentary records. These factors may decouple the Zr/Rb ratio from actual grain size, potentially introducing uncertainties when interpreting paleoflood events solely based on this proxy. In addition, in fluvial core sediments, the particle density and surface roughness have a stronger influence on Rb than on Ti, which reduces the accuracy of the Zr/Rb ratio as a grain size proxy [28]. Therefore, in paleohydrological studies, grain size proxies should be carefully tested or calibrated against measured grain size parameters before being applied.

5. Discussion

5.1. Reconstruction of Paleo-Hydrological Sedimentary Sequences

Insitu XRF core scanning technology (X-ray fluorescence) is based on the principle of X-ray excitation, whereby the cumulative X-ray fluorescence signals of elements are used to characterize elemental abundances, allowing for the acquisition of continuous, non-destructive, and millimeter-scale high-resolution geochemical signals [30]. In fluvial~lacustrine sediments, high moisture content can significantly affect the XRF scanning intensity of light elements [28]. This is mainly because the water content of the core sediments exceeds 60%, forming a thin water film between the Ultralene foil and the sediment surface, thereby weakening the scanning intensity of light elements such as Al and Si. In contrast, heavier elements such as Zr, Ti, and Rb release stronger energy under X-ray excitation and are less affected by the physical properties of the sediments, thus more accurately reflecting their concentration variations [26]. In addition, to eliminate the influence of core elemental intensity noise and background values from the watershed, this study utilized R 4.3.3 software and Peakfit 38 software, applying the local polynomial regression method (span = 0.2) to fit the logarithmic elemental ratios and coarse-grain end-member component curves. Since the paleoflood sequences reconstructed from alluvial plains differ from seasonal flood events and primarily focus on high-magnitude flood events, peaks exceeding a threshold were identified as paleoflood events [1,6] (Figure 6). Based on this, the paleoflood sequence for the desert section of the Yellow River was reconstructed from the past 30 ka (Figure 6).Due to the disparity in sampling resolution between the grain-size analyses (5 cm) and the XRF-derived elemental data (2 mm), the identification of paleo-flood events based on grain-size end-member modeling and elemental proxies reveals minor inconsistencies. For example, during the 2.5~4 ka period, the paleoflood sequence reconstructed using the ln(Zr/Ti) elemental ratio displays multiple peaks, while the sequence reconstructed using the grain size end-member component method shows no peaks. Therefore, the chemical elemental-based paleoflood events are relatively more complete. According to the peak values defined by the threshold, paleoflood in the desert section of the Yellow River primarily occurred between 30~12 ka and 3~0 ka. From 30 to 12 ka, the frequency and intensity of paleoflood were relatively high, while during the Holocene, the frequency and intensity of these floods decreased. There were almost no large paleoflood records during the early to mid-Holocene, while in the late Holocene, both the frequency and intensity of paleoflood increased significantly.

5.2. Paleoflood Evolution and Its Response to the Climate

To explore the relationship between paleoflood and climate, this study compares the paleoflood sequence in the desert section of the upper Yellow River with regional environmental indicators (Figure 5). Among these, Greenland ice core δ18O changes indicate global temperature variations [38], total inorganic carbon (TIC) content in Daihai sediment cores reflects regional temperature changes [34], and δ18O values from stalagmites in Dongge Cave (Guizhou) [35], Hulu Cave (Nanjing) [37], and the Summer Monsoon Index (SMI) of Qinghai Lake [36] reflect precipitation changes under the influence of the East Asian summer monsoon. During the 30~12 ka period, the Qinghai Lake summer monsoon index was low, magnetic susceptibility in the Loess Plateau of Longxi was at low values, and the δ18O values from Hulu Cave and Dongge Cave stalagmites were high, all indicating a weakened summer monsoon and drier climate. Additionally, Greenland ice core δ18O values were low, and ice-wedge sand in the Ordos Plateau suggested a temperature decrease of about 13 °C compared to present day, further indicating cooler conditions. This dry and cold climate, coupled with reduced vegetation cover, enhanced slope erosion in the watershed, leading to increased surface runoff and sediment entering the river, resulting in frequent high-magnitude paleoflood.
In the early to mid-Holocene, the frequency and intensity of paleoflood decreased compared to the Last Glacial Maximum (LGM). Palynological records show that the vegetation was dominated by Artemisia and Stipa species [39]. Sediment records from the Ulan Buh Desert and the northern desert regions [40] along with lower clay content in Daihai sediments [41], also suggest a drier environment. During the early to mid-Holocene, the climate was warm and wet, with better vegetation development in the watershed. The riverbanks’ resistance to erosion improved, leading to decreased runoff and sediment input to the river, and a more even distribution of precipitation throughout the year, resulting in lower paleoflood frequency and intensity. Therefore, the fewer paleoflood records from the early to mid-Holocene likely responded to the warmer and wetter climate conditions. In the late Holocene, the frequency and intensity of paleoflood significantly increased. Palynological data from NingwuGonghai in Shanxi show a sharp decrease in tree pollen, with grass pollen dominating [42], and the total inorganic salt content in Daihai sediments in Inner Mongolia reached its lowest level in the Holocene [34], both indicating a more arid climate. Additionally, sediment analysis from Huyang Lake in Shaanxi shows coarser grain size, lower total organic carbon, and CaCO3 content, as well as a lower lake level, suggesting cooler and drier conditions [43]. Furthermore, the temperature reconstruction for the Holocene from environmental change literature shows relatively low temperatures during the 3~0 ka period [44]. Paleoflood studies from the middle Yellow River [21,45] and rivers in the southwestern United States [46] strongly support the conclusion that paleofloods were more frequent under cool and dry climate conditions. Overall, the frequency and intensity of paleofloods during the dry and cold periods of the desert section of the Yellow River were greater than during the warm and wet periods, with the frequency and intensity being lowest during the warm and wet phase.

6. Conclusions

This study focuses on the sediment core HDZ04 from the desert section of the upper Yellow River, analyzing the sediment grain size and elemental characteristics to preliminarily establish a record of paleoflood in this region over the past 30 ka. By combining regional environmental indicators, the study explores the response relationship between paleoflood and climate change. Using the EMMA end-member model, four end-member components were extracted, with the percentage of two coarser end-members selected as substitute indicators of flood dynamics. Pearson correlation analysis of the ln(Zr/Ti), ln(Zr/Rb) values, and the composite grain size indicator revealed that ln(Zr/Ti) had a more significant correlation with the composite grain size value, so Zr/Ti was chosen as the elemental proxy indicator for paleoflood reconstruction. By integrating physical and chemical indicators and OSL dating results, this study reconstructed the paleoflood sequence of the desert section of the Yellow River over the past 30,000 years. The results show that from 30 to 12 ka, paleoflood frequency and intensity were high, while in the Holocene, the frequency and intensity of paleoflood decreased. During the early and middle Holocene, there was a slight increase in paleoflood frequency and intensity, while in the late Holocene (3~0 ka), the frequency and intensity of paleoflood significantly increased. Comparison with regional climate indicators revealed that under cold and dry climate conditions, the frequency and intensity of paleoflood in the desert section of the Yellow River increased. It should be noted that in this study, the ln(Zr/Ti) peak value of Yellow River core sediments was used to reflect flood hydrodynamic conditions. Due to the resolution of the study scale, these peaks may reflect the intensity of a single paleoflood event or the accumulation of several large flood deposits. Therefore, considering the potential for stratigraphic gaps caused by factors such as river meandering in alluvial plain areas, future research should obtain continuous core data to more completely reconstruct the paleoflood record.

Supplementary Materials

The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/atmos16091019/s1.

Author Contributions

Conceptualization, H.P.; methodology, Y.J.; validation, H.P. and Y.J.; writing—review and editing, H.P. and Y.J. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the National Science Foundation of China (No. 42001004, 42001008 and 42402194; the Science and Technology Achievements Transformation Project of Inner Mongolia Autonomous Region (2021CG0046); the Science and Technology Plan Project of the Alxa League (AMYY 2021-19).

Data Availability Statement

Some datasets are publicly available in the Supplementary Materials. Additional datasets, including raw XRF scanning files and sediment core measurements, are available from the corresponding author upon reasonable request.

Acknowledgments

We sincerely appreciate the constructive feedback provided by the three anonymous reviewers.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Toonen, W.; Winkels, T.; Cohen, K.; Prins, M.; Middelkoop, H. Lower Rhine historical flood magnitudes of the last 450 years reproduced from grain-size measurements of flood deposits using End Member Modelling. Catena 2015, 130, 69–81. [Google Scholar] [CrossRef]
  2. Baker, V.R. Paleoflood hydrology: Origin, progress, prospects. Geomorphology 2008, 101, 1–13. [Google Scholar] [CrossRef]
  3. Baker, V.R. Paleoflood hydrology and extraordinary flood events. J. Hydrol. 1987, 96, 79–99. [Google Scholar] [CrossRef]
  4. Huang, C.C.; Pang, J.; Zha, X.; Zhou, Y.; Su, H.; Zhang, Y.; Wang, H.; Gu, H. Holocene palaeoflood events recorded by slackwater deposits along the lower Jinghe River valley, middle Yellow River basin, China. J. Quat. Sci. 2012, 27, 485–493. [Google Scholar] [CrossRef]
  5. Mao, P.; Pang, J.; Huang, C.; Zha, X.; Zhou, Y.; Guo, Y.; Zhou, L. A multi-index analysis of the extraordinary paleoflood events recorded by slackwater deposits in the Yunxi Reach of the upper Hanjiang River, China. Catena 2016, 145, 1–14. [Google Scholar] [CrossRef]
  6. Jones, A.F.; Macklin, M.G.; Brewer, P.A. A geochemical record of flooding on the upper River Severn, UK, during the last 3750 years. Geomorphology 2012, 179, 89–105. [Google Scholar] [CrossRef]
  7. Gilli, A.; Anselmetti, F.S.; Glur, L.; Wirth, S.B. Lake sediments as archives of recurrence rates and intensities of past flood events. In Dating Torrential Processes on Fans and Cones; Springer: Dordrecht, The Netherlands, 2013; pp. 225–242. [Google Scholar]
  8. Ielpi, A.; Ghinassi, M. A sedimentary model for early Palaeozoic fluvial fans, Alderney Sandstone Formation (Channel Islands, UK). Sediment. Geol. 2016, 342, 31–46. [Google Scholar] [CrossRef]
  9. Agbotui, P.Y.; Firouzbehi, F.; Medici, G. Review of Effective Porosity in Sandstone Aquifers: Insights for Representation of Contaminant Transport. Sustainability 2025, 17, 6469. [Google Scholar] [CrossRef]
  10. Giguet-Covex, C.; Arnaud, F.; Enters, D.; Poulenard, J.; Millet, L.; Francus, P.; David, F.; Rey, P.-J.; Wilhelm, B.; Delannoy, J.-J. Frequency and intensity of high-altitude floods over the last 3.5 ka in northwestern French Alps (Lake Anterne). Quat. Res. 2012, 77, 12–22. [Google Scholar] [CrossRef]
  11. Yang, J.; Liu, Z.; Yin, J.; Tang, L.; Zhao, H.; Song, L.; Zhang, P. Paleoflood reconstruction in the lower Yellow River floodplain (China) based on sediment grain size and chemical composition. Water 2023, 15, 4268. [Google Scholar] [CrossRef]
  12. Lim, J.; Lee, J.; Hong, S.; Kim, J. Late Holocene flooding records from the floodplain deposits of the Yugu River, South Korea. Geomorphology 2013, 180–181, 109–119. [Google Scholar] [CrossRef]
  13. Peng, F.; Kasse, C.; Prins, M.A.; Ellenkamp, R.; Krasnoperov, M.Y.; van Balen, R.T. Paleoflooding reconstruction from Holocene levee deposits in the lower Meuse valley, The Netherlands. Geomorphology 2021, 352, 107002–107014. [Google Scholar] [CrossRef]
  14. Schillereff, D.N.; Chiverrell, R.C.; Macdonald, N.; Hooke, J.M. Flood stratigraphies in lake sediments: A review. Earth-Sci. Rev. 2014, 135, 17–37. [Google Scholar] [CrossRef]
  15. Schulte, L.; Peña, J.C.; Carvalho, F.; Schmidt, T.; Julià, R.; Llorca, J.; Veit, H. A 2600 year history of floods in the Bernese Alps, Switzerland: Frequencies, mechanisms and climate forcing. Hydrol. Earth Syst. Sci. Discuss. 2015, 12, 3391–3448. [Google Scholar] [CrossRef]
  16. Pang, H.L.; Jia, Y.X.; Li, F.Q.; Qin, L.; Chen, L.Y. An improved method for paleoflood reconstruction from core sediments in the upper Yellow River. Front. Earth Sci. 2023, 11, 1149502. [Google Scholar] [CrossRef]
  17. Munoz, S.E.; Giosan, L.; Therrell, M.D.; Remo, J.W.F.; Shen, Z.; Sullivan, R.M.; Wiman, C.; O’dOnnell, M.; Donnelly, J.P. Climatic control of Mississippi River flood hazard amplified by river engineering. Nature 2018, 556, 95–98. [Google Scholar] [CrossRef]
  18. Hagstrom, C.A.; Leckie, D.A.; Smith, M.G. Point Bar Sedimentation and Erosion Produced by an Extreme Flood in a Sand and Gravel-Bed Meandering River. Sediment. Geol. 2018, 377, 1–16. [Google Scholar] [CrossRef]
  19. Yu, S.-Y.; Hou, Z.; Chen, X.; Wang, Y.; Song, Y.; Gao, M.; Pan, J.; Sun, M.; Fang, H.; Han, J.; et al. Extreme flooding of the lower Yellow River near the Northgrippian-Meghalayan boundary: Evidence from the Shilipu archaeological site in southwestern Shandong Province, China. Geomorphology 2020, 350, 106878. [Google Scholar] [CrossRef]
  20. Zhang, Y.; Huang, C.C.; Pang, J.; Zha, X.; Zhou, Y.; Wang, X. Holocene palaeoflood events recorded by slackwater deposits along the middle Beiluohe River valley, middle Yellow River basin, China. Boreas 2015, 44, 127–138. [Google Scholar] [CrossRef]
  21. Fan, L.; Huang, C.C.; Pang, J.; Zha, X.; Zhou, Y.; Li, X.; Liu, T. Sedimentary records of palaeofloods in the Wubu reach along the Jin-Shaan gorges of the middle Yellow River, China. Quat. Int. 2015, 380, 368–376. [Google Scholar] [CrossRef]
  22. Li, B.; Ge, Q.S. Evolution of the Yellow River in the Hetao Plain of Inner Mongolia in the past 2000 years. Acta Geogr. 2003, 58, 239–246. [Google Scholar]
  23. Pan, B.; Pang, H.; Zhang, D.; Guan, Q.; Wang, L.; Li, F.; Guan, W.; Cai, A.; Sun, X. Sediment grain-size characteristics and its source implication in the Ningxia–Inner Mongolia sections on the upper reaches of the Yellow River. Geomorphology 2015, 246, 255–262. [Google Scholar] [CrossRef]
  24. Yang, G.S.; Ta, W.Q.; Dai, F.N.; Liu, Y.X.; Jing, K.; Li, B.Y.; Zhang, O.Y.; Lu, R.; Hu, L.F.; Tao, Y. Contribution of sand sources to the silting of riverbed in inner Mongolia section of Yellow river. J. Desert Res. 2003, 23, 152–159, (In Chinese with English abstract). [Google Scholar]
  25. Yao, Z.; Ta, W.; Jia, X.; Xiao, J. Bank erosion and accretion along the Ningxia–Inner Mongolia reaches of the Yellow River from 1958 to 2008. Geomorphology 2011, 127, 99–106. [Google Scholar] [CrossRef]
  26. Richter, T.O.; van der Gaast, S.; Koster, B.; Vaars, A.; Gieles, R.; de Stigter, H.C.; De Haas, H.; van Weering, T.C.E. The Avaatech XRF core scanner: Technical description and applications to NE Atlantic sediments. Geol. Soc. Lond. Spéc. Publ. 2006, 267, 39–50. [Google Scholar] [CrossRef]
  27. Löwemark, L.; Chen, H.-F.; Yang, T.-N.; Kylander, M.; Yu, E.-F.; Hsu, Y.-W.; Lee, T.-Q.; Song, S.-R.; Jarvis, S. Normalizing XRF-scanner data: A cautionary note on the interpretation of high-resolution records from organic-rich lakes. J. Asian Earth Sci. 2011, 40, 1250–1256. [Google Scholar] [CrossRef]
  28. Pang, H.L.; Gao, H.S.; Liu, X.P.; Tian, W.Q.; Zou, Y.; Pan, B.T. Preliminary study on calibration of X-ray fluorescence core scanner for quantitative element records in the Yellow River sediments. Quat. Sci. 2016, 36, 237–246, (In Chinese with English abstract). [Google Scholar]
  29. Fralick, P.W.; Kronberg, B.I. Geochemical discrimination of clastic sedimentary rock sources. Sediment. Geol. 1997, 113, 111–124. [Google Scholar] [CrossRef]
  30. Kylander, M.E.; Ampel, L.; Wohlfarth, B.; Veres, D. High-resolution X-ray fluorescence core scanning analysis of Les Echets (France) sedimentary sequence: New insights from chemical proxies. J. Quat. Sci. 2011, 26, 109–117. [Google Scholar] [CrossRef]
  31. Chen, J.; Chen, Y.; Liu, L.; Ji, J.; Balsam, W.; Sun, Y.; Lu, H. Zr/Rb ratio in the Chinese loess sequences and its implication for changes in the East Asian winter monsoon strength. Geochim. Cosmochim. Acta 2006, 70, 1471–1482. [Google Scholar] [CrossRef]
  32. Dypvik, H.; Harris, N.B. Geochemical facies analysis of fine-grained siliciclastics using Th/U, Zr/Rb and (Zr+Rb)/Sr ratios. Chem. Geol. 2001, 181, 131–146. [Google Scholar] [CrossRef]
  33. Weltje, G.J.; Tjallingii, R. Calibration of XRF core scanners for quantitative geochemical logging of sediment cores: Theory and application. Earth Planet. Sci. Lett. 2008, 274, 423–438. [Google Scholar] [CrossRef]
  34. Xiao, J.; Wu, J.; Si, B.; Liang, W.; Nakamura, T.; Liu, B.; Inouchi, Y. Holocene climate changes in the monsoon/arid transition reflected by carbon concentration in Daihai Lake of Inner Mongolia. Holocene 2006, 16, 551–560. [Google Scholar] [CrossRef]
  35. Dykoski, C.; Edwards, R.; Cheng, H.; Yuan, D.; Cai, Y.; Zhang, M.; Lin, Y.; Qing, J.; An, Z.; Revenaugh, J. A high-resolution, absolute-dated Holocene and deglacial Asian monsoon record from Dongge Cave, China. Earth Planet. Sci. Lett. 2005, 233, 71–86. [Google Scholar] [CrossRef]
  36. An, Z.; Colman, S.M.; Zhou, W.; Li, X.; Brown, E.T.; Jull, A.J.T.; Cai, Y.; Huang, Y.; Lu, X.; Chang, H.; et al. Interplay between the Westerlies and Asian monsoon recorded in Lake Qinghai sediments since 32 ka. Sci. Rep. 2012, 2, 619. [Google Scholar] [CrossRef] [PubMed]
  37. Wang, Y.J.; Cheng, H.; Edwards, R.L.; An, Z.S.; Wu, J.Y.; Shen, C.-C.; Dorale, J.A. A High-Resolution Absolute-Dated Late Pleistocene Monsoon Record from Hulu Cave, China. Science 2001, 294, 2345–2348. [Google Scholar] [CrossRef]
  38. Rasmussen, S.O.; Andersen, K.K.; Svensson, A.M.; Steffensen, J.P.; Vinther, B.M.; Clausen, H.B.; Siggaard-Andersen, M.; Johnsen, S.J.; Larsen, L.B.; Dahl-Jensen, D.; et al. A new Greenland ice core chronology for the last glacial termination. J. Geophys. Res. Atmos. 2006, 111, 907–923. [Google Scholar] [CrossRef]
  39. Huang, F. Holocene grassland vegetation, climate and human impact in central eastern Inner Mongolia. Sci. China 2005, 48, 1025–1039. [Google Scholar] [CrossRef]
  40. Chen, F.; Li, G.; Zhao, H.; Jin, M.; Chen, X.; Fan, Y.; Liu, X.; Wu, D.; Madsen, D. Landscape evolution of the Ulan Buh Desert in northern China during the late Quaternary. Quat. Res. 2014, 81, 476–487. [Google Scholar] [CrossRef]
  41. Peng, Y.; Xiao, J.; Nakamura, T.; Liu, B.; Inouchi, Y. Holocene East Asian monsoonal precipitation pattern revealed by grain-size distribution of core sediments of Daihai Lake in Inner Mongolia of north-central China. Earth Planet. Sci. Lett. 2005, 233, 467–479. [Google Scholar] [CrossRef]
  42. Chen, F.; Xu, Q.; Chen, J.; Birks, H.J.B.; Liu, J.; Zhang, S.; Jin, L.; An, C.; Telford, R.J.; Cao, X.; et al. East Asian summer monsoon precipitation variability since the last deglaciation. Sci. Rep. 2015, 5, 11186. [Google Scholar] [CrossRef] [PubMed]
  43. Yan, Y.; Zhou, J.; He, Z.; Sun, Q.; Fei, J.; Zhou, X.; Zhao, K.; Yang, L.; Long, H.; Zheng, H. Evolution of Luyang Lake since the last 34,000 years: Climatic changes and anthropogenic impacts. Quat. Int. 2016, 440, 90–98. [Google Scholar] [CrossRef]
  44. Hou, G.; Fang, X. Characteristics of Holocene Temperature Change in China. Prog. Geogr. 2011, 30, 1075–1080. [Google Scholar]
  45. Huang, C.C.; Pang, J.; Zha, X.; Su, H.; Jia, Y.; Zhu, Y. Impact of monsoonal climatic change on Holocene overbank flooding along Sushui River, middle reach of the Yellow River, China. Quat. Sci. Rev. 2007, 26, 2247–2264. [Google Scholar] [CrossRef]
  46. Harden, T.; Macklin, M.G.; Baker, V.R. Holocene flood histories in south-western USA. Earth Surf. Process. Landf. 2010, 35, 707–716. [Google Scholar] [CrossRef]
Atmosphere 16 01019 g001
Figure 1. Geographical location of core and grain sizecomposition of core sediments.
Figure 1. Geographical location of core and grain sizecomposition of core sediments.
Atmosphere 16 01019 g001
Atmosphere 16 01019 g002
Figure 2. Grain size composition core sediments and characteristics of chemical elements are analyzed by in-situ scanning.
Figure 2. Grain size composition core sediments and characteristics of chemical elements are analyzed by in-situ scanning.
Atmosphere 16 01019 g002
Atmosphere 16 01019 g003
Figure 3. Results of grain size End-Member Modeling.
Figure 3. Results of grain size End-Member Modeling.
Atmosphere 16 01019 g003
Atmosphere 16 01019 g004
Figure 4. Fitted grain size curve. The shaded area represents the error range, and the dashed line indicates the background value.
Figure 4. Fitted grain size curve. The shaded area represents the error range, and the dashed line indicates the background value.
Atmosphere 16 01019 g004
Atmosphere 16 01019 g005
Figure 5. Comparison of grain size indicators and geochemical element indicators of core sediments.
Figure 5. Comparison of grain size indicators and geochemical element indicators of core sediments.
Atmosphere 16 01019 g005
Atmosphere 16 01019 g006
Figure 6. Comparison between multi-proxy-derived paleoflood records and regional climate indicators. (a) The ln(Zr/Ti) values exceeding the threshold (this study); (b) fitted grain size values above the threshold (this study); (c) the total inorganic carbon in Daihai Lake [34]; (d) δ18O values from stalagmites in Dongge Cave [35]; (e) the Summer Monsoon Index of Qinghai Lake [36]; (f) δ18O values from stalagmites in Hulu Cave [37].
Figure 6. Comparison between multi-proxy-derived paleoflood records and regional climate indicators. (a) The ln(Zr/Ti) values exceeding the threshold (this study); (b) fitted grain size values above the threshold (this study); (c) the total inorganic carbon in Daihai Lake [34]; (d) δ18O values from stalagmites in Dongge Cave [35]; (e) the Summer Monsoon Index of Qinghai Lake [36]; (f) δ18O values from stalagmites in Hulu Cave [37].
Atmosphere 16 01019 g006
Table 1. OSL dating results of sediments from core HDZ04.
Table 1. OSL dating results of sediments from core HDZ04.
Depth
(m)		K
(%)		Th
(ppm)		U
(ppm)		Water Content (%)Dose Rate
(Ga/ka)		De
(Gy)		Age
(ka)
6.11.54 ± 0.078.51 ± 0.261.89 ± 0.1422 ± 52.2 ± 0.173.01 ± 0.211.4 ± 0.1
14.21.47 ± 0.065.01 ± 0.191.18 ± 0.1315 ± 51.94 ± 0.1651.14 ± 1.4926.3 ± 2.3
19.81.53 ± 0.064.47 ± 0.181.09 ± 0.1321 ± 51.72 ± 0.1451.21 ± 1.5329.7 ± 2.5
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Pang, H.; Jia, Y. Sedimentary Records of Paleoflood Events in the Desert Section of the Upper Yellow River Since the Late Quaternary. Atmosphere 2025, 16, 1019. https://doi.org/10.3390/atmos16091019

AMA Style

Pang H, Jia Y. Sedimentary Records of Paleoflood Events in the Desert Section of the Upper Yellow River Since the Late Quaternary. Atmosphere. 2025; 16(9):1019. https://doi.org/10.3390/atmos16091019

Chicago/Turabian Style

Pang, Hongli, and Yunxia Jia. 2025. "Sedimentary Records of Paleoflood Events in the Desert Section of the Upper Yellow River Since the Late Quaternary" Atmosphere 16, no. 9: 1019. https://doi.org/10.3390/atmos16091019

APA Style

Pang, H., & Jia, Y. (2025). Sedimentary Records of Paleoflood Events in the Desert Section of the Upper Yellow River Since the Late Quaternary. Atmosphere, 16(9), 1019. https://doi.org/10.3390/atmos16091019

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop