• Aucun résultat trouvé

Hydrological behaviour of a small tropical catchment on volcanic deposits

N/A
N/A
Protected

Academic year: 2021

Partager "Hydrological behaviour of a small tropical catchment on volcanic deposits"

Copied!
1
0
0

Texte intégral

(1)

Laboratoire des Interactions Sol-Agrosystème-Hydrosystème UMR LISAH, 2 place Pierre Viala, F-34060 Montpellier cedex 1 - France

http://sol.ensam.inra.fr/Lisah/

Systèmes de cultures bananes, plantains et ananas

CIRAD

UR Systèmes de cultures bananes, plantains et ananas, Neufchâteau, Sainte-Marie, F-97130 Capesterre Belle-Eau - France http://www.cirad.fr/

Volcanic insular reliefs, as the Lesser Antilles arc, are locally subject to a strong anthropic pressure. These regions are characterized by abundant rainfall, a strong heterogeneity of the geometry of the deposits and a high soil infiltration capacity.

¾ identification of hydrological

processes at the catchment scale

Acknowledgements This work is funded by the Guadeloupe Region (FWI), the Ministère de l’Ecologie et du Développement Durable (France) and the European Community. The authors are very grateful to Germain Onapin and Colbert Béhary for their active contribution on the field.

Hydrological behaviour of a small tropical catchment on volcanic deposits

Hydrological behaviour of a small tropical catchment on volcanic deposits

Jean-Baptiste Charlier

1,

*, Roger Moussa

2

, Philippe Cattan

1

et Marc Voltz

2

1CIRAD, UPR 26, Station de Neufchâteau, Sainte-Marie, 97130 Capesterre Belle Eau, France ; 2INRA, UMR LISAH, Bat. 24, 2 place Viala, 34060 Montpellier Cedex 1, France

* Author : Tel.: +33 590 86 29 98; fax: +33 590 86 80 77; e-mail address : [email protected]

• surface = 19.5 ha mean slope = 12 % altitude = 350 m • annual rainfall : 4230 mm (2003) to 7030 mm (2004) • Hydrological measurements during 2 years (2003 & 2004)

1. Objectives

5. Global model approach

4. Hydrological behaviour

¾shallow aquifer in piled ash soils layer: high infiltration capacity (Ks) in umbric andosols; drainage by the stream and percolation to the deep aquifer; thickness ≈ 6 m

¾deep aquifer in the nuées ardentes layer: medium hydraulic conductivity (K); the underlying

weathered breccia layer represent the substratum; thickness = 30 m

Scale in meters

Ks = 8e-06m/s K = 8.e-07m/s

Hydrogeologic diagram of Féfé catchment

3. Hydrological processes

Stormflow events are intense and rapid

• minimum response time = 25 min

runoff coefficient = 6 to 24 %and is correlated to the rainfall volume and the initial soil water humidity

Runoff process at the evenly time scale

Baseflow is considered like an indicator of the initial soil water humidity of the catchment Measured Calculated Streamflow volume [m3] peakflow [L/s] 4 602 4 975 163 847 154 074 Nash coef. [-] Streamflow Log10(streamflow) S1 0.82 S2 0.83 0.81 0.91

6. Conclusion & perspectives

¾ Runoff production function :

hortonian overlandflow

¾ Transfer function :

diffusive wave model (Hayami, 1951)

¾ Drawdown of water table :

exponential function Infiltration Transfer function Stock 1 Shallo w aquifer Rainfall ET Baseflow Surface runoff Stock 2 Deep aquifer Streamflow Deep percolation

Global conceptual model of Féfé catchment

Reference Hayami S., 1951. On the propagation of flood waves. Disaster Prev. Res. Inst. Bull. Université de Kyoto, Japon,1: 1-16.

Example of simulation : the excessively wet month may 2004

¾ The main hydrological process is deep percolation; the rapid and intense surface runoff during storm is supposed to be dominated by overland flow; the shallow aquifer is contributing to streamflow (baseflow).

¾ A global conceptual model based on a diagram of the hydrological behaviour of Féfé catchment validate the advanced hypothesis related to water budget components.

¾ In order to characterize the spatial variability of the hydrological processes due to the heterogeneity of the volcanic deposits, a spatial modelling approach is request for next studies.

¾ High rainfall intensities ¾ fast and intense response of the stream to a rainfall event ¾ fast and low shallow groundwater variations ¾ smooth and high deep

groundwater variations

¾ high performance of the global model approach

¾ highlight problems related to the simulation of low flow events

2 aquifer systems in the volcanic deposits

Surface runoff Recession (47 %) Baseflow Stormflow (53 %) Streamflow Evapo-transpiration (3.4 mm/day) Deep aquifer (nuées ardentes) ΔS = 0 Rainfall Shallow aquifer (piled ash soils)

ΔS = 0 1303 (31 %) 390 (9 %) 1886 (45 %) 3840 (81 %) 161 (4 %) Baseflow 489 (11 %) Fluxes (mm) (%) year 2003 4230 (100 %) Legend

¾Data : Rainfall, evapotranspiration (ET), saturated hydraulic conductivity (Ks)

¾Calibration : fit of exponential function Characteristics of storm event (18/05/2004) :

2. The Féfé experimental catchment, Guadeloupe (FWI)

Ks

¾ the main water balance

components is deep percolation

Références

Documents relatifs

With the application of deep learning techniques to computer vision domain, the choice of appropriate loss function for a task has become a critical aspect of the model training..

The parameters of interest are spatial variability and correlation length of permeabil- ity, absolute permeability, pore compressibility, cap rock permeability, and

Comparison of stock market prediction by using the base dataset of stock price and then a combination of the stock price with additional external parameters of crude

Con- cerning particle and bacteria transport, results demonstrate that (1) a first turbidity signal occurs during increasing discharge due to remobilization of particles from

Improvement comes with bottles that are only opened at desired depth, or with bottles having an initial volume equal to zero that is enlarged through the functioning of a moving

To represent the linear fractional order system of (10) by a linear time-invariant system model so as to derive their closed form impulse and step responses, its irrational

To represent the linear fractional order system of (10) by a linear time-invariant system model so as to derive their closed form impulse and step responses,

Bilmes, “Acoustic classification using semi-supervised deep neural networks and stochastic entropy- regularization over nearest-neighbor graphs,” in IEEE Interna- tional Conference