Elsevier

Ecological Informatics

Volume 64, September 2021, 101363
Ecological Informatics

Reconstruction of springs discharge using tree-rings and earlywood vessel chronologies in an alluvial aquifer

https://doi.org/10.1016/j.ecoinf.2021.101363Get rights and content

Highlights

  • Tree-rings and vessel features were used to reconstruct the past springs discharge.

  • The vessel diameter and the tree-ring widths were the best inputs for the ANN model.

  • Dendrohydrology and ANN showed a high performance in the modeling process.

  • The best modeling results were observed in medium and minimum discharges.

Abstract

Monitoring changes in groundwater storage is necessary for water resource management plans. The temporal records of spring discharge are restricted by the absence or paucity in many parts of the world. To address this shortcoming, various models and methods have been implemented to reconstruct past hydrologic data. In this study, temporal variations of spring discharge were reconstructed using tree-rings and earlywood vessel chronologies in an alluvial aquifer on the southern coast of the Caspian sea. The dendrochronological analysis was performed using both cross-sections and core samples of Zelkova carpinifolia (Pall.) K. Koch species, including tree-rings width and vessel features (vessels diameter, area, and perimeter) as proxy data. The distance between the sampling trees varied between 20 and 1000 m. An artificial neural network (ANN) was used to establish a relationship between the temporal dendrochronological records of Z. carpinifolia and corresponding discharge obtained from two alluvial springs in the vicinity of sampling trees. The ANN model was optimized and tested by comparing the recorded and predicted values in the training and testing stage of the modeling process. Results show the high performance of the ANN model in the training (R-squared = 0.98, MSE = 0.009) and test stage (R-squared = 0.81, MSE = 1.26). The results also indicated the higher correlation of springs discharge with vessel features (R = 0.76) compared to tree-rings (R = 0.59). Dendrohydrology works better for those springs with notable discharge changes during the growing seasons. Further, the combination of tree-ring width and vessel features as inputs in modeling will increase the model's performance. The tested model was applied to the dendrochronological records of past decades (1982 to 2001) to reconstruct the spring discharge data during the growing seasons for those years.

Introduction

Springs are a vital source of water for many local communities around the world. During the past decades, population growth, climatic changes, and over-exploitation of groundwater have exerted extra pressure on these valuable resources, especially in the arid and semi-arid regions (Green et al., 2011; Jyrkama and Sykesa, 2007). Spring is a natural outpouring of the groundwater, and it occurs when the water table reaches the ground surface. Springs can be divided into two main types based on geologic structures: alluvial and karst springs. Alluvial springs are found in the alluvial sediments and plains. Karst springs are found in the karst formations located in mountainous areas. The majority of springs in Iran are categorized as Karst springs. However, most alluvial springs are located in densely populated areas and, therefore, they are prone to over-extraction and water pollution (Gholami et al., 2008). Groundwater or springs are one of the main sources of soil moisture, significantly influencing trees that grow in the surrounding environments around water bodies (Mazza et al., 2020; Tulik et al. 2020).

Studies show that groundwater resources have been declining in many parts of the world (Wada et al., 2010). Therefore, it is necessary to record the spring discharge for better management of groundwater resources. Monitoring changes in groundwater is necessary for adopting strategies to mitigate the impacts of water deficiency and climate change. Temporal variation in groundwater depth is tipically monitored on local scales using water table measurements from piezometric wells. Unfortunately, such practices are hampered in many parts of the world by the absence or scarcity of monitoring stations, temporal and spatial gaps in the collected time series data, and data quality issues. Even for the areas with available records, the data commonly span less than one century. Therefore, different models and techniques have been developed to reconstruct past groundwater storage or spring discharge records.

Dendrohydrology is the science of reconstructing past hydrologic variabilities from tree-rings and vessel features (Anderson et al., 2019; Blake et al., 2020; Gholami et al., 2017; Gholami et al., 2019; Tsvetanov et al., 2020; Wigley et al., 1984). Tree-rings and vessel features (dendrohydrology) records have been successfully applied to the reconstruction of rivers discharge and flood occurrence (Hardman and Reil, 1936; Hawley, 1937; Woodhouse, 2000; Pederson et al., 2001; Woodhouse and Meko, 2002; Liu et al., 2010; Edison et al., 2013; Gholami et al., 2017; Anderson et al., 2019; Meko and Therell, 2019; Zhang et al., 2020a, Zhang et al., 2020b). Previous studies have shown that precipitation, soil moisture, soil depth, altitude, and temperature affect tree-ring width, and therefore, these parameters can be investigated using dendrochronological studies (Briffa et al., 1988; Conkey, 1979; Gray et al., 2007; Hidalgo et al., 2000; Meko et al., 2001; Naurzbaev and Sidorova, 2004; Oie et al., 2020; Wise, 2010; Woodhouse et al., 2010; Yin et al., 2008).

Bogino and Jobbágy (2011) investigated the relationship between climate and groundwater fluctuations by studying the growth and death of the Prosopis caldenia species in the lowlands and uplands. They concluded that groundwater depth has a vital role in trees' lives and that the optimum depth of groundwater was ranging from 2 to 8 m. In fact, in areas where trees source their water need from groundwater, they can suffer groundwater level lowering due to climatic change or human impacts because it also affects the water content stored in the upper soil layers (Lewis and Burgy, 1964; Mazza et al., 2020; Williams et al., 1998). Gholami et al. (2017) simulated groundwater level fluctuations using tree-ring and vessel features in the north of Iran. They concluded that vessel features represent the groundwater fluctuations even better than tree-rings. Anderson et al. (2019) successfully reconstructed the streamflow data for the Tennessee Valley using tree-rings chronology. Meko and Therell (2019) used tree-ring and vessel widths for studying floods on the White River of Arkansas, providing results on their potential for precisely modeling the maximum river flows. Previous studies used regression analysis to reconstruct past hydrologic parameters based on tree-rings and vessel features (Robertson et al., 1990; Liu, 2011; Gholami et al., 2017; Tulik et al. 2020). Vessel features are closely related to water absorption and conduction. Therefore, this can be a suitable parameter for evaluating the changes in precipitation, groundwater level, and spring discharge in the alluvial aquifers because there is a correlation between earlywood vessel features and soil moisture during the growing seasons (Garcia Gonzalez and Eckstein, 2003; Fonti and Garcia Gonzalez, 2004; Gholami et al., 2015 and Gholami et al., 2017; Kames et al., 2016; Tulik et al. 2020). In the alluvial aquifers, the water saturates the pore spaces of aquifers. Therefore, the groundwater depth and springs discharge are dependent on groundwater recharge, precipitation, and the amount of groundwater extraction. All the mentioned factors affect the spring discharge fluctuations.

The tree-rings width and the vessel's features vary by environmental factors during the growing season. The wood produced at the beginning of the growing season is called earlywood or springwood and is characterized by coarse vessels and a thin wall. The wood produced in the late growing season is called the latewood or summerwood, and it can be identified by its relatively small vessels and thick walls. The combination of earlywood and latewood forms the annual tree-ring. At the beginning of the growing season, water absorption in trees increases due to increased precipitation, spring discharge, groundwater level, and increased vessel size (Campelo et al., 2010). The change in vessel size is attributed to water availability, where larger vessels have larger space for water transportation (Gholami et al., 2017; Hacke and Sperry, 2001). Therefore, tree-rings and vessel features can be used in the spring discharge modeling as an alternative in measuring groundwater and spring discharge.

Caucasian elm (a binomial and ring-porous tree) is a mesophytic deciduous species that grows mainly in the riverside, mixed lowland, and ravine forests. These species are suitable for evaluating the effects of soil moisture and groundwater changes during the growing season and have been used for dendrohydrology in previous studies (Davis et al., 2012; Gholami et al., 2017; Gholami et al., 2019; Grissino-Mayer, 1993). Previous dendrohydrology studies were mainly focused on surface water (e.g., reconstruction of river discharge, flood occurrence). To our knowledge, dendrochronology has not yet been used in the reconstruction of spring discharge. It is assumed that.

by identifying the relationship between the spring discharge and the tree-ring widths and vessel features, we can reconstruct the historical spring discharge (Gholami et al., 2017; Mazza et al., 2020; Williams et al., 1998). In this research, a new approach for reconstructing spring discharge is introduced. For this purpose, tree-rings, vessel features, and hydrogeologic variables were used to reconstruct the discharge of two alluvial springs during the three temporal sections (early spring, mid-growing season, and late-summer) of the growing season in the past decades. This study aimed to predict the historical spring's discharge using dendrohydrology and ANN.

Section snippets

Study area

The study area is a portion of the southern coastal plains of the Caspian Sea in Iran, extending between 48° 59′ E and 37° 39′ N (Fig. 1). The mean annual precipitation is approximately 1000 mm, which creates a humid climate in the region. Major land-use types in the study area are forest lands, rice field lands, and residential areas. The lithology of the site is mainly Quaternary sediments (alluvial sediments) that host alluvial springs. This study focused on two springs within the study site

Results

The monthly discharge during the growing season (from 1982 to 2018) ranged from 1.7 to 11.8 l per second for Vardem spring and 2 to 20 l per second for the Khanekenar spring. The mean monthly springs discharge during the growing season in Vardem and Khanekenar springs are 4.1 and 6.4 lit per second, respectively. Both of the studied springs have permanent discharge in the growing season. The mean monthly vessel diameters vary from 47 to 232 μm, and the mean tree-rings width range from 1.4 to

Discussion

The number of vessels did not show a significant relationship with the springs' discharge. We can not be sure that all of the small vessels were counted carefully. Therefore, the number of vessels is not a suitable parameter compared to the other vessel features (Gholami et al., 2019; Kames et al., 2016; Tulik et al. 2020). Also, earlywood vessels' size decreases significantly during the growing season. The maximum and minimum vessel size was observed in the early spring and late summer,

Conclusions

The results show that dendrohydrology can be used for reconstructing the past springs' discharge in the unconfined aquifers within the alluvial formations. The performance of the dendrohydrology-based spring discharge modeling is dependent on environmental conditions such as the distance from rivers, the aquifer's properties, the tree species, and the discharge regime of the springs.

Dendrohydrology performs better for those springs that the flow fluctuates a lot during growing seasons. It is

Declaration of Competing Interest

None.

Acknowledgments

We would like to express our appreciation to the Regional Water Company of Guilan for providing the spring discharge data and helping us with the data preprocessing.

References (63)

  • S. Kames et al.

    Continuous earlywood vessels chronologies in floodplain ring-porous species can improve dendrohydrological reconstructions of spring high flows and flood depths

    J. Hydrol.

    (2016)
  • R. Kumar et al.

    Comparison of regression and artificial neural network models for estimation of global solar radiations

    Renew. Sust. Energ. Rev.

    (2015)
  • H. Liu

    Impact of climate change on groundwater recharge in dry areas: an ecohydrology approach

    J. Hydrol.

    (2011)
  • Y. Liu et al.

    Tree-ring hydrologic reconstructions for the Heihe River watershed, western China since AD 1430

    Water Res.

    (2010)
  • H. Sahour et al.

    A comparative analysis of statistical and machine learning techniques for mapping the spatial distribution of groundwater salinity in a coastal aquifer

    J. Hydrol.

    (2020)
  • N. Samani et al.

    A simple neural network model for the determination of aquifer parameters

    J. Hydrol.

    (2007)
  • N. Tsvetanov et al.

    First measurements of blue intensity from Pinus peuce and Pinus heldreichii tree rings and potential for climate reconstructions

    Dendrochronologia

    (2020)
  • T. Zhang et al.

    Tree-ring width based streamflow reconstruction for the Kaidu River originating from the central Tianshan Mountains since A.D. 1700

    Dendrochronologia

    (2020)
  • F. Alshehri et al.

    Mapping the distribution of shallow groundwater occurrences using remote sensing-based statistical modeling over Southwest Saudi Arabia

    Remote Sens.

    (2020)
  • F. Anctil et al.

    Evaluation of neural networks streamflow forecasting on 47 watersheds

    J. Hydrol. Eng. ASCE

    (2005)
  • A. Blake et al.

    Palaeoclimate potential of New Zealand Manoao colensoi (silver pine) tree rings using Blue-Intensity (BI)

    Dendrochronologia

    (2020)
  • F. Campelo et al.

    Vessel features of Quercus ilex L. growing under Mediterranean climate have a better climatic signal than tree-ring width

    Trees.

    (2010)
  • L.E. Conkey

    Response of tree-ring density to climate in Maine, U.S.A

    Tree-Ring Bull.

    (1979)
  • E.R. Cook et al.

    The smoothing spline: a new approach to standardizing forest interior tree-ring width series for dendroclimatic studies

    Tree-Ring Bull.

    (1981)
  • E.L. Davis et al.

    Evaluating the suitability of nine shelterbelt species for dendrochronological purposes in the Canadian Prairies

    Agrofor. Syst.

    (2012)
  • E. Edison et al.

    Vessel anomalies in Quercus macrocarpa tree rings associated with recent floods along the Red River of the North, United States

    Water Resour. Res.

    (2013)
  • P. Fonti et al.

    Suitability of chestnut earlywood vessel chronologies for ecological studies

    New Phytol.

    (2004)
  • P. Fonti et al.

    Studying global change through investigation of the plastic responses of xylem anatomy in tree rings

    New Phytol.

    (2009)
  • H.C. Fritts

    Tree Rings and Climate

    (1976)
  • H.C. Fritts

    Tree Rings and Climate

    (2001)
  • I. Garcia Gonzalez et al.

    Climatic signal of earlywood vessels of oak on a maritime site

    Tree Physiol.

    (2003)
  • Cited by (0)

    View full text