Elsevier

Journal of Hydrology

Volume 550, July 2017, Pages 386-398
Journal of Hydrology

Research papers
Wavelet analysis of remote sensing and discharge data for understanding vertical ground movements in sandy and clayey terrains of the Po Delta area (Northern Italy)

https://doi.org/10.1016/j.jhydrol.2017.05.017Get rights and content

Highlights

  • Wavelet analyses of geodetic and hydrological data are used for studying subsidence.

  • CGPS data positively/negatively correlate with river discharge data.

  • At the inter-annual scale river bank subsidence rates depend on geology and wet-dry lap.

  • At the intra-annual scale river bank subsidence trends depend on river discharge.

Abstract

Subsidence phenomena change the proneness of urban and coastal areas to huge flooding, particularly, in river delta regions. Many natural and/or anthropic processes can induce vertical ground movements; identifying the causes of the observed phenomena is a useful, even if not easy, task for flood forecasting. In order to improve the knowledge of natural versus anthropic contributions to vertical ground movements, in the present work data analysis techniques (polynomial regression analysis, frequency analysis and cross wavelet and wavelet coherence analysis) are used to analyze both the long- and short-time scale dependencies between Continuous Global Positioning System (CGPS) and variations of meteorological, hydrogeological and hydrological data collected in the Po River Delta area (Northern Italy) from January 2009 to December 2015. The main findings are as follows: (i) occurrence of positive correlations and periodic oscillations of about 2–3 years for all the analyzed meteorological, hydrogeological and hydrological records; (ii) a general decrease of the ground level with a faster soil lowering in clayey terrain with respect to sandy terrain; and, (iii) negative correlations between CGPS and Po River discharge records at the intra-annual scale (2–6 months), which are attributed to an isostatic rebound induced by the change of the river discharge due to precipitation changes over northern Italy.

Introduction

Over the past 25 years the number of large floods in Europe has increased (Kundzewicz et al., 2013), likely due to intensification of rainfalls and inadequacy of river morphologies and geometries to contain the higher peak-flow discharges. The significant human and economic loss caused by vulnerability of the Italian landscape to flooding events (e.g. Po Plain, 1951: 84 victims; Firenze, 1966: 35 victims) has stressed the importance to establish cause-effect relationships underlying the observed phenomena. For example, at least eight flooding episodes with daily peak discharge above 8000 m3/s occurred, since 1916 specifically in 1917, 1926, 1928, 1951, 1976, 1994, 2000 and 2014, with the absolute maximum discharge observed on 20 May 1926 (9780 m3/s) (Zanchettin et al., 2008a).

Undoubtedly, flooding processes are favored by downward ground movements, but understanding the causes of these movements is not an easy task due to the high number of variables involved in the natural and anthropic processes and their complex interrelationships. Subsidence of urbanized alluvial plains is studied (Galloway and Burbey, 2011) by comparing the ground vertical movements, measured by using geodetic techniques such as high precision levelling or satellite methods, against the ground water level of aquifers, the extracted water volume and the rainfall. This comparison is often based on qualitative approaches or basic statistical tools, such as linear regression applied to publicly available low resolution data, typically monthly means (e.g., Osmanoğlu et al., 2011, Cigna et al., 2012). More advanced analyses are achieved when stratigraphic information, soil mechanical properties and aquifer system data are available, as they allow the development of hydrodynamic flow and soil compaction models (e.g., Sneed and Galloway, 2000, Calderhead et al., 2011).

Differently from the above mentioned studies, in this paper we propose a multi-scale statistical analysis to identify potential relationships between variations of meteorological, hydrogeological and hydrological data and high-resolution remote sensing data (CGPS). The proposed approach is applied to data acquired in the Po Delta area (Northern Italy) (Fig. 1), an area subject to both, natural and anthropic subsidence. Since the end of 1800, measurements of land subsidence started to be collected over the Po Delta area and, from the second half of 20th century, these measures have been attributed to both natural and anthropic causes (Gambolati and Teatini, 1998, Carminati and Di Donato, 1999, Carminati and Martinelli, 2002). Anthropogenic factors have been significantly reduced since 1970, when the intensive phase of methane extraction ceased, although in some areas it is still present (Fabris et al., 2014).

In particular, two complementary approaches of data analysis are proposed for retrieving information at inter- and intra-annual time scales in the period ranging from 2009 to 2015. Specifically, polynomial regression, histogram analysis and cross correlation of annual cumulative values are used for studying the inter-annual scale up to a period of 5 years, while Cross Wavelet Transform (CWT) and Wavelet Transform Coherence (WTC) analyses (Grinsted et al., 2004) are used for the intra-annual time scale. The former approach highlights the meteorological forcing on hydrological and hydrogeological systems, while the latter allows to better determine the physical mechanisms, such as infiltration and isostatic processes, which likely drive the ground deformations during shorter periods.

Section snippets

Study area

The Po Plain (inset in Fig. 1) represents the foreland basin of southern Alps and northern Apennines (Doglioni, 1993), two mountain chains generated during the Tertiary by the collision between the African and European continental plates. The thrusting phase brought to the production of a considerable syn-orogenic clastic sequence and to the infilling of the foredeep basins. During the Plio-Quaternary, a sediment layer of more than 7 km of thickness accumulated on the southern side of the Po

Datasets

Data used in this work have been collected from 10 stations located in the Po Delta area (see Fig. 1) and grouped based on the different nature (i.e., hydrological, hydrogeological, meteorological and geodetic) of the analyzed time series records. All data are available on open source websites or on request by the producing institutions. Although the data coming from each station are inhomogeneous in terms of temporal sampling step and coverage, the performed analyses allowed to define the main

Data analysis

In order to identify and characterize possible common features among the considered datasets (see Section 3), two complementary approaches are proposed:

  • (1)

    Polynomial regression and histogram analysis of annual cumulative values to investigate the long-time scale (multi-year) dependences;

  • (2)

    Cross wavelet transform (CWT) and wavelet transform coherence (WTC) analysis to investigate the short-time scale (weekly to-monthly) dependences.

Discussion

In this section, we discuss the most relevant results and provide hypotheses about the processes responsible of the observed trends. In addition, some hypotheses on the mechanisms that control the river-bank movements in the Po Delta area are also proposed with reference to the different soil's nature of the study area.

Conclusions

Two complementary approaches for the analysis of meteorological, hydrological, hydrogeological and geodetic data have been proposed to study the vertical ground movements of the Po Delta area (Northern Italy). The first approach, based on polynomial regression and histogram analysis of cumulative values, allowed to retrieve information at inter-annual time scales; the second approach, based on cross wavelet transform (CWT) and wavelet transform coherence (WTC) analyses, was used for retrieving

Acknowledgments

The authors acknowledge support of the research Project: “High Resolution Geomatic Methodologies for Monitoring Subsidence and Coastal Changes in the Po Delta area” (n. CPDA158140, 2015) funded by the University of Padua, Italy. Two reviewers are also acknowledged for their comments and suggestions, which helped to improve and clarify the original manuscript.

References (39)

  • A. Amorosi et al.

    Facies architecture and Latest Pleistocene-Holocene depositional history of the Po Delta (Comacchio area), Italy

    J. Geol.

    (2003)
  • M.P. Baldwin et al.

    The quasi-biennial oscillation

    Rev. Geophys.

    (2001)
  • M. Bondesan et al.

    Coastal areas at risk from storm surges and sea-level rise in north-eastern Italy

    J. Coastal Res.

    (1995)
  • L. Bonzi et al.

    Coastal unconfined aquifer of Emilia Romagna: geological model and salinization state

    Il Geologo dell'Emilia-Romagna

    (2010)
  • Burrato, P., Ciucci, F., Valensise, G., 2003. An inventory of river anomalies in the Po Plain, Northern Italy: evidence...
  • E. Carminati et al.

    Separating natural and anthropogenic vertical movements in fast subsiding areas: the Po plain (N. Italy) case

    Geophys. Res. L

    (1999)
  • E. Carminati et al.

    The fragile balance of the Po Valley

    Le Scienze

    (2006)
  • Cibin, U., Stefani, M., 2009. Notes of the Italian Geological Map at scale 1:50.000, Foglio 187, Codigoro. Servizio...
  • A. Correggiari et al.

    Depositional patterns in the Late-Holocene Po delta system

  • Cited by (4)

    • Benthic foraminifers from Holocene subaqueous deltas of the Western Mediterranean: Stratigraphic implications and palaeoenvironmental significance of the biofacies

      2021, Marine Geology
      Citation Excerpt :

      On a regional scale, the Po Plain-Adriatic Sea system belongs to the Alpine-Apennine and Dinarides-Hellenides foreland, formed in response to the convergence between the European and African plates (Doglioni, 1993; Boccaletti and Martelli, 2011). Subsidence rates in the delta region due to tectonic subsidence plus sediment compaction equal ~3–4 mm/yr (Vitagliano et al., 2017). Among the active Po river outlets (Fig. 1A), the Pila branch makes the highest water (~60% of the total) and sediment discharge values (~75%), but other (palaeo)branches were alternatively dominant in the past (Correggiari et al., 2005).

    View full text