Navigation – Plan du site

Experimental study of post-eruption overland flow and sediment load from slopes overlain by pyroclastic-flow deposits, Unzen volcano, Japan

Étude expérimentale du ruissellement et de sa charge solide sur les pentes du volcan Unzen (Japon) couvertes par des dépôts pyroclastiques
Yasuhiro Ogawa, Hiromu Daimaru et Akira Shimizu
p. 237-246

Résumés

Les volcans actifs représentent des zones de recherche privilégiées pour qui veut mesurer les processus de ruissellement et d’entraînement des particules, en comparant les pentes recouvertes de débris pyroclastiques avec celles restées indemnes. Lors des éruptions de 1990 à 1995, d’épais dépôts pyroclastiques émis par le volcan Unzen ont recouvert ses pentes, entraînant le déclenchement de l’érosion dont les manifestations ont rapidement diminué une fois l’éruption terminée. Pour mieux comprendre ce processus de diminution de la mobilité des particules peu après la fin de l’éruption, le ruissellement et sa capacité de transport ont été mesurés sur deux parcelles expérimentales : l’une avait été ensemencée artificiellement, l’autre présentait un sol nu qui, pour l’expérimentation, fut recouvert d’un voile de cendres volcaniques fines. Le rapport entre la charge solide transportée par le ruissellement et le volume total d’eau ruisselée sur les parcelles est un indice (S/Q) qui varie selon le degré de stabilisation du sol à la surface des dépôts pyroclastiques. Il a varié d’une part avec la nature des matériaux en surface (apport du voile de cendres sur sol nu) et d’autre part avec la croissance du couvert végétal. Deux situations expliquent la diminution de la mobilité des particules sur les pentes inscrites dans les dépôts pyroclastiques après la fin de l’éruption du volcan Unzen. Dans un cas, l’érosion du voile de cendres volcaniques fines, artificiellement saupoudré sur l’un des sites, a rétabli la capacité d’infiltration du sol dépourvu de végétation et a diminué le ruissellement ; dans l’autre c’est la colonisation par les plantes qui, après l’éruption, a rendu la surface du sol plus stable.

Haut de page

Errata

Article soumis le 18 mars 2006, accepté le 31 août 2007.

Notes de la rédaction

Acknowledgments
The authors thank the members of the Kyushu Regional Office of the Forestry Agency and the Shimabara Branch Office of the Nagasaki Prefectural Government for their cooperation in establishing and monitoring the field sites. We also thank Y. Miyabuchi, T. Shimizu, I. Noda, M. Ideta and T. Sassa of the Forestry and Forest Products Research Institute for their help in installing the monitoring devices and the experimental volcanic ash layer. H. Ochiai, T. Sanmori and Y. Adikari of the Forestry and Forest Products Research Institute are thanked for valuable advice on preparing this paper. The authors are grateful to J.-C. Thouret, D. Palacios, J.A. Palmer, S. J. Cronin, and J. Raffy for reviewing our manuscript. K. Martelli has improved the English grammar, and J.-C. Thouret and J. Raffy have helped us in translating English into French.

Texte intégral

Introduction

1Unzen volcano, a volcanic dome complex located in southwestern Japan, began erupting in 1990 after 198 years of quiescence. Eruptions lasting 5 years produced over 9,400 pyroclastic flows, and together with numerous debris flows (produced during heavy rainfall; Harada et al., 1999), caused a large volume of sediment to accumulate at the periphery of the lava dome (Miyabuchi, 1999). Thick layers of volcanic material were deposited on the formerly forested volcano's slopes. By May 1995, the total volume of pyroclastic-flow deposits had reached 120 million m3 (Ishikawa et al., 1996).

2Studies of sediment load on the flanks of active volcanoes such as Mount Usu, Mount St. Helens, Mount Galunggung and Mount Merapi, have shown that sediment load and erosion from slopes (on which a number of debris flows and mudflows have been deposited) drop sharply when volcanic eruptions cease (Yamamoto, 1984; Collins and Dunne, 1986; Pierson and Janda, 1992; Jitousono et al., 1996).

3The Kyushu Regional Forest Office and the Shimabara Branch Office of the Nagasaki Prefectural Government undertook aerial seeding at Mount Unzen, in order to reduce erosion from the surface of the pyroclastic-flow deposits after recent volcanic activity (Nagasaki Prefectural Office and Nagasaki Prefectural Shimabara Branch Office, 2000). Nishida et al. (1998a) and Yamakoshi and Suwa (1998) monitored sediment dynamics at sites on pyroclastic-flow deposits and reported that sediment load decreased after the revegetation of slopes, even where vegetation was recent and immature. They did not report on the dynamics of sediment movement during and soon after deposition of the volcaniclastic material. The processes of sediment load from slopes covered by recent pyroclastic-flow deposits during and immediately after volcanic eruptions remains to be understood.

4It has been suggested that the ground surface conditions (i.e. surface soil structure, grain-size distribution, and permeability) of pyroclastic-flow deposits differ with time elapsed since eruption (Nishida et al., 1996; Nishida and Mizuyama, 1998b). Adikari et al. (2005) suggested that vegetation cover plays an important role in post-eruption surface erosion processes on Mount Usu. Sediment load and surface erosion on pyroclastic flow deposits before and after an eruption clearly differ, due to the effects of the tephra mantle and vegetation, which change the hydrological conditions of the substrate.

5This study seeks to clarify the effects of tephra mantle and vegetation on the overland flow and sediment load from bare and revegetated slopes of Unzen volcano (figs.1 and 2).

Fig. 1 – Location of field sites.
Fig. 1 – Emplacement des sites d’étude.

Fig. 1 – Location of field sites.Fig. 1 – Emplacement des sites d’étude.

Fig. 2 – Oblique aerial view of part of the Unzen volcano in 1997, from the southeast. Photograph was taken by the Kyushu Regional Forest Office.
Fig. 2 – Depuis le sud-est, vue aérienne oblique d’une partie du volcan Unzen en 1997. Photo prise par le Bureau Forestier Régional de Kyushu.

Fig. 2 – Oblique aerial view of part of the Unzen volcano in 1997, from the southeast. Photograph was taken by the Kyushu Regional Forest Office.Fig. 2 – Depuis le sud-est, vue aérienne oblique d’une partie du volcan Unzen en 1997. Photo prise par le Bureau Forestier Régional de Kyushu.

BS: bare site; RS: revegetated site.
BS : site dénudé ; RS : site revégétalisé.

6This is achieved with a combination of experimental research and field observations. The experimental work reproduces the ground conditions that are prevalent in fine-grained volcanic ash both immediately following the eruption and four years later. Field data were collected on overland flow, sediment load, and vegetation cover of the volcano.

Field sites, survey and experimental methods

7Two contrasting field sites were selected in the Akamatsudani basin on pyroclastic material deposited between 1991 and 1995 (tab. 1). Both field sites have a concave slope and a rill.

Table 1 – Characteristics of the field sites.
Tableau 1 – Caractéristiques des sites d’étude.

Field sites

Drainage area (m2)

Mean slope angle (degrees)

Vegetation species

Bare site

220

11

sparse vegetation (Lb, Pc)

Revegetated site

2759

15

Lb, Ec, Lc, Pc, Ms

Lb: Lespedeza bicolor var. Japonica Nakai; Ec: Eragrostis curvula Ness; Lc: Lespedeza cuneata G. Don; Ms: Miscanthus sinensis Anderss; Pc: polygonum cuspidatum Sieb et Zucc.

8Pyroclastic deposits in the area are between 10 and 20 m thick (Ishikawa et al., 1996), with boulders up to 6 m in diameter littering the ground. The "bare" site (32°44’56” N, 130°19’26” E; 335 m) was characterised by sparse vegetation. The "revegetated" site (32°44’54” N, 130°19’11” E; 361 m) was located in an area where the Kyushu Regional Forest Office had conducted aerial seeding in April 1996. Seeds of eleven plant species, suspended in fertilizer solution, were sprayed from a helicopter. There was a minute litter layer at the revegetated site during the observation interval. One woody plant species and four herbaceous plant species were present at the revegetated site; one of the four herbaceous plant species present was not among the seeds spread in 1996. The ground surface deposit of both field sites had similar grain-size distribution and hydraulic conductivity. At the onset of the field survey in 1998, little ash fall deposit was left on the ground surface of the field sites. The grain-size distribution of the ground surface deposit ( <4 cm deep) sampled at the bare site consists of clay (0.73%), silt (3.78%), sand (51.04%), and gravel (44.45%). Cobbles (mostly <50 cm) and boulders were also present in the ground surface deposit but not sampled for grain-size distribution. Hydraulic conductivity of the ground surface deposit was determined to be 5.38 x 10-3 cm/s-1 in 1998 (Ogawa et al., 2006).

Monitoring overland flow and sediment load at observation sites

9A 0.5-mm tipping-bucket rain gauge was installed at the base of the experimental sites (fig. 3).

Fig. 3 – Topographic map and equipment at the bare site.
Fig. 3 – Carte topographique et équipement du site dépourvu de végétation.

Fig. 3 – Topographic map and equipment at the bare site.Fig. 3 – Carte topographique et équipement du site dépourvu de végétation.

1: rain gauge; 2: weir and sediment trap; 3: sieved ash treatment area; 4: plate dividing watershed; 5: gully dividing watershed; 6: natural topographic boundary to the watershed; 7: water flow in a heavy rainfall event; 8: surface soil monitoring area.
1 : pluviomètre ; 2 : déversoir et piège à sédiments ; 3 : aire saupoudrée de cendres tamisées ; 4 : tôle délimitant l’aire contributive du ruissellement ; 5 : ravine délimitant une aire contributive du ruissellement ; 6 : limite naturelle de bassin versant ; 7 : cheminement de l’eau en cas de fortes pluies ; 8 : aire de la parcelle expérimentale.

10A sediment trap (1.8 m long, 0.44 m wide, and 0.5 m deep) and a triangular notch weir were installed at the downstream end of each observation site. A plate was installed to prevent extraneous surface water from entering the bare site (solid line in fig. 3). Overland flow was determined by converting the water level – monitored at one-minute intervals using a hydraulic water gauge installed in the triangular notch weir – into discharge (sediment load). Volume and grain-size distributions of trapped sediment were measured for material that accumulated in the sediment trap. Observations of overland flow and sediment load were carried out on the bare and the revegetated sites. They were monitored at the revegetated site from 27 August 1998 to 14 December 2001 while the monitoring at the bare site ended on 16 November 1999, due to construction (tab. 2). Rainfall monitoring resumed at the revegetated site on 29 October 1999.

Table 2 – Treatment on the bare site using sieved ash and observation schedule between 1998 and 1999.
Tableau 2 – L’apport des cendres tamisées sur la parcelle dépourvue de végétation et le calendrier des observations de 1998 à 1999.

Date

Bare site

Revegetated site

27 August 1998

Start monitoring rainfall, overland flow, sediment load

Start monitoring overland flow, sediment load

11 June 1999

Sieved ash treatment

29 October 1999

Stop monitoring rainfall

Resume monitoring rainfall

16 November 1999

Stop monitoring overland flow, sediment load

14 December 1999

Stop monitoring overland flow, sediment load

11On 11 June 1999, fine-grained volcanic ash was spread over about one-third of the area of the bare site (68 m2; fig. 3). This ash material had been collected in 1993 from the summit area of Mount Iwatoko about 2 km southwest of the lava dome, and sieved through a 0.42 mm mesh to remove dust and coarse particles. The fine-grained volcanic ash with which the ground surface was artificially veneered is termed “sieved ash”. The grain-size distribution of the sieved ash consisted of clay (10.50%), silt (63.14%), and sand (26.36%). The objective of the sieved ash treatment was to simulate the accumulation of ash that fell on the bare site during an eruption. All vegetation in the 68 m2-experimental area was mowed, and the ground was then covered with a total of 135.1 kg of sieved ash, which amounted to 0.2 and 0.5 cm in thickness. Four soil pits were studied to monitor movement of the ash in the ash treatment area. Movement of the ash during rainfall was documented using sketches and photographs, while profiles of the ground surface were recorded at intervals of one to three weeks. The ash was not spread over the revegetated site during the monitoring period in order to better distinguish the influence of the vegetation.

Vegetation survey of the revegetated site

12Surveys of coverage ratio, growth height and coverage were conducted each summer at the revegetated site. Coverage ratio is defined as the percentage of ground covered by vegetation in a 1 m2 survey area. Grass cover was assessed using a line transect method, in which two 10 m-long survey lines were established, the first across the slope (east-west) and the second down the slope (north-south). The cumulative percentage of line length covered by each species was recorded. Basal coverage was measured for herbaceous plants, and crown coverage and tree height were assessed for woody plants. Vegetation was also surveyed annually from 1996 to 1999 about 50 m upstream of the site (Nagasaki Prefectural Office and Nagasaki Prefectural Shimabara Branch Office, 2000). In order to compare the changes in vegetation cover following the 1996 aerial seeding, we used data obtained from the two field sites as well as from the survey carried out 50 m upstream during this study.

Results

Site monitoring prior to sieved ash treatment

13Hydrographs of the two study locations were monitored on 26 May 1999 prior to the sieved ash treatment (fig. 4).

Fig. 4 – Hydrograph of overland flow at the field sites. Hydrograph of the bare site before the sieved ash treatment.
Fig. 4 – Hydrogramme du ruissellement sur les sites d’étude. L’hydrogramme du site dénudé a trait à la période précédant l’apport de cendres tamisées.

Fig. 4 – Hydrograph of overland flow at the field sites. Hydrograph of the bare site before the sieved ash treatment.Fig. 4 – Hydrogramme du ruissellement sur les sites d’étude. L’hydrogramme du site dénudé a trait à la période précédant l’apport de cendres tamisées.

14Peak overland flow was 0.14 mm/min at the revegetated site but was as high as 0.44 mm/min at the bare site. An overland flow event that took place on 26 May 1999 at the bare site from 22:28 h to 22:55 h (local time) was not observed at the revegetated site. Peak overland flow and total sediment load were smaller at the revegetated site than at the bare site.

15Total overland flow at the revegetated site and the bare site during the observation interval, including the overland flow event on 26 May 1999, was 3.05 mm and 12.63 mm respectively. Sediment load per unit area from the revegetated site and the bare site were 0.26 g/m2 and 1.60 g/m2 respectively. Total overland flow from the revegetated site was about four times lower than that from the bare site. Sediment load removed from the revegetated site was about six times less than that from the bare site.

Site monitoring following the sieved ash treatment, and subsequent vegetation changes

16Temporal changes in sediment load are expressed using the ratio of overland flow (Q, in mm) to trapped sediment load (S, in kg). The ratio (i.e. S/Q, both measured in kg/mm) is referred to as “Sb/Qb” for the bare site and “Sr/Qr” for the revegetated site

Measures at the bare site

17Sb/Qb at the bare site was 0.03 and 0.08 before the sieved ash treatment, and rising to 1.41 in mid-June 1999 immediately following the treatment (fig. 5).

Fig. 5 – Temporal changes in total rainfall (R), overland flow (Qb and Qr), sediment load, over-land flow ratio (Sb/Qb and Sr/Qr) and median sediment clast diameter (d50) collected in the sediment traps at the two field sites.
Fig. 5 – Variabilité des précipitations totales (R), du ruissellement (Qb et Qr), de la charge solide, du rapport entre charge solide et ruissellement (Sb/Qb et Sr/Qr) et de la médiane du diamètre (d50) des particules récoltées dans les pièges à sédiments sur les deux sites d’étude.

Fig. 5 – Temporal changes in total rainfall (R), overland flow (Qb and Qr), sediment load, over-land flow ratio (Sb/Qb and Sr/Qr) and median sediment clast diameter (d50) collected in the sediment traps at the two field sites.Fig. 5 – Variabilité des précipitations totales (R), du ruissellement (Qb et Qr), de la charge solide, du rapport entre charge solide et ruissellement (Sb/Qb et Sr/Qr) et de la médiane du diamètre (d50) des particules récoltées dans les pièges à sédiments sur les deux sites d’étude.

18In late June the value dropped to between 0.49 and 1.07 and in late September dropped to approximately 0.30 (which was still greater than before the ash treatment). During a storm on 11 September, two debris flows were initiated in a gully upstream of the field sites. The total rain and overland flow during the period August 31 to September 23 (including the debris flows of 11 September) was greater than other periods. The Sb/Qb ratio decreased following this period.

19Immediately after the sieved ash treatment (between 11 and 21 June 1999), the median diameters (d50) of trapped sediment particles (fig. 5) was smaller than that after 21 June, and was similar to the median diameter of the original sieved ash (d50: 0.04 mm). Sieved ash was removed from the site even by low overland flow, of 0.32 mm, between 11 and 21 June 1999. The total volume of trapped sediment removed between 11 and 21 June 1999 was 0.45 kg.

20Overland flow before and after the sieved ash treatment was then compared during two overland flow events resulting from similar rainfall conditions: before treatment on 7 June 1999 and after treatment on 24 June 1999 (tab. 3).

Table 3 – Hydrological conditions at the bare site on 7 June 1999 and 24 June 1999.
Tableau 3 – Conditions hydrologiques sur le site dépourvu de végétation le 7 juin 1999 et le 24 juin 1999.

7 June

24 June

Rainfall (mm)

18.0

20.5

30 min. antecedent rainfall (mm)

8.5

9.5

20 min.antecedent rainfall (mm)

6.5

9.5

5 min. antecedent rainfall (mm)

3.0

3.0

Maximum 5 min. rainfall (mm)

6.0

3.5

Maximum 10 min. rainfall(mm)

8.5

6.5

Duration of overland flow (min.)

48

55

Peak flow (mm/min.)

0.33

0.46

Total overland flow volume (mm)

2.80

5.51

Overland flow coefficient (%)

15.6

26.9

Overland flow coefficient is the ratio of overland flow amount (mm) versus total rainfall (mm) during each overland flow event.
Le coefficient de ruissellement est le rapport entre la quantité d’eau ruisselée (mm) par rapport au total des précipitations (mm) durant chaque épisode pluvieux.

21The maximum 5 and 10 minute rainfall intensities during overland flow on 24 June were only 58% and 76% of the amounts on 7 June but the peak overland flow and the total overland flow on 24 June were, respectively, 1.4 times and two times greater than those on 7 June. Peak overland flow followed peak rainfall and occurred more quickly after, than before, the sieved ash treatment (fig. 6).

Fig. 6 – Hydrograph of overland flow at bare site before (A) and after (B) the sieved ash treatment.
Fig. 6 – Hydrogramme du ruissellement sur la parcelle nue avant (A) et après (B) le saupoudrage de cendres tamisées.

Fig. 6 – Hydrograph of overland flow at bare site before (A) and after (B) the sieved ash treatment.Fig. 6 – Hydrogramme du ruissellement sur la parcelle nue avant (A) et après (B) le saupoudrage de cendres tamisées.

22Temporal changes in the ground surface at the bare site were documented between one day prior to and about one month after the sieved ash treatment (fig. 7).

Fig. 7 – View and surficial sediment profile of the sieved ash treatment area. A: view before the ash treatment; B: view immediately after the ash treatment; C: soil profile immediately after the ash treatment; D: soil profile one month after the ash treatment; E: view one month after the ash treatment.
Fig. 7 – Vue et coupes des sédiments superficiels de la parcelle ayant été saupoudrée de cendres tamisées. A : vue avant l’apport des cendres ; B : vue immédiatement après le saupoudrage des cendres ; C : profil du sol immédiatement après l’apport des cendres ; D : profil du sol un mois après le saupoudrage des cendres ; E : vue un mois après l’apport des cendres.

Fig. 7 – View and surficial sediment profile of the sieved ash treatment area. A: view before the ash treatment; B: view immediately after the ash treatment; C: soil profile immediately after the ash treatment; D: soil profile one month after the ash treatment; E: view one month after the ash treatment.Fig. 7 – Vue et coupes des sédiments superficiels de la parcelle ayant été saupoudrée de cendres tamisées. A : vue avant l’apport des cendres ; B : vue immédiatement après le saupoudrage des cendres ; C : profil du sol immédiatement après l’apport des cendres ; D : profil du sol un mois après le saupoudrage des cendres ; E : vue un mois après l’apport des cendres.

Pr: surface profile monitoring area; Sa: sieved ash; G: gravel layer; SaG: sieved ash within gravel layer; P: pyroclastic-flow deposit beneath gravel layer.
Pr : parcelle étudiée ; Sa : cendres tamisées ; G : couche de graviers ; SaG : cendres tamisées infiltrées dans la couche de gravier ; P : dépôt de coulée pyroclastique sous la couche de graviers.

23Sieved ash was evident immediately after the treatment (fig. 7B, Pr “surface profile of the monitoring area”, and 7C, Sa “sieved ash”) but was absent on the ground surface one month later (fig. 7E). Sieved ash was, however, detected one month after treatment in the subsurface at a depth of 0.7 to 1.5 cm in the interstices of a gravel layer (fig. 7D, SaG). Three months after the treatment, no sieved ash was found in the near-surface layers (SaG) where it had previously been observed.

Measures at the revegetated site

24Annual rainfall was highest in 1999 (tab. 4; fig. 5), for example the annual rainfall was approximately 1.3 times greater than it was in 2000 and 2001.

Table 4 – Annual observation results at the revegetated site.
Tableau 4 – Résultats de l’observation annuelle du site revégétalisé.

Year

Rainfall
(mm/year)

Overland flow
(mm/year)

Sediment load
(kg/year)

1998

291.0

0

0

1999

2672.5

70.23

8.00

2000

2074.5

13.67

0.19

2001

1964.5

0.41

0

1998 data stem from observations made only between 28 August and 31 December because no observation was carried out before 28 August.
Les données de 1998 ont uniquement trait à la période du 28 août au 31 décembre, car aucune observation n’a été conduite auparavant.

25The annual rainfall has been measured at the revegetated site since 2001. Annual rainfall in 2003 and 2004 was 2535 mm and 2613.5 mm, respectively. While the rainfall in 1999 was not abnormal, the annual sediment load was highest at 8 kg. Total sediment load in 2000 was only 0.19 kg; 2.6% of the load that was removed in 1999. Sediment load was no more detectable in 2001. Sr/Qr was 0.03 to 0.23 in 1999, decreasing to as little as 0.004 to 0.02 in 2000. In 2001, overland flow took place with little sediment load (tab. 4; fig. 5). At the revegetated site, Sr/Qr gradually decreased regardless of the overland flow during the monitoring period.

26Vegetation at the site was assessed one month before initiation of the experiment (July 1998; fig. 8A), and approximately three years later (fig. 8B).

Fig. 8 – Changes in mean growth height, coverage ratio and vegetation type at the revegetated site. A: view in July, 1998; B: view in August, 2001. The 1996-1999 data are from the Shimabara branch office of Nagasaki prefectural government (2000) and the 2000-2001 data is from the authors.
Fig. 8 – Accroissement de la taille moyenne des végétaux, du taux de couverture et du type de végétation sur le site revégétalisé. A : vue en juillet 1998 ; B : vue en août 2001. Les données de 1996 à 1999 proviennent du bureau de Shimabara de la préfecture de Nagasaki (2000), tandis que celles de 2000 à 2001 proviennent des auteurs.

Fig. 8 – Changes in mean growth height, coverage ratio and vegetation type at the revegetated site. A: view in July, 1998; B: view in August, 2001. The 1996-1999 data are from the Shimabara branch office of Nagasaki prefectural government (2000) and the 2000-2001 data is from the authors.Fig. 8 – Accroissement de la taille moyenne des végétaux, du taux de couverture et du type de végétation sur le site revégétalisé. A : vue en juillet 1998 ; B : vue en août 2001. Les données de 1996 à 1999 proviennent du bureau de Shimabara de la préfecture de Nagasaki (2000), tandis que celles de 2000 à 2001 proviennent des auteurs.

27The majority of the dark green spots in fig. 8 are that of a woody plant that grew during the period of observation. The coverage ratio (measured in 1 m2 areas) increased markedly from 1996 to 1997, having reached 100% since 1997. The mean height of herbaceous and woody plants increased from 1996 to 1999, reflecting steady plant growth. In the 2000-2001 survey, the mean height of herbaceous plants was about 1 m and 2.1 to 2.3 m for woody plants (Fig. 8C). The coverage data reflects the growth of branches (i.e. crown coverage) of the woody plants. Vegetation at the site was not a simple community of herbaceous plants but rather consisted of a tiered structure with an upper layer of 1 to 2 m-high woody plants, as well as intermediate and lower layers of herbaceous plants. The coverage by herbaceous plants in August 2000 and September 2001 was 46.1% and 39.2% and the coverage by woody plants was 32.4% and 45.3%, respectively (Fig. 8C).

28The median diameters of the sediment particles at the revegetated site were measured in 1999 and 2000. There was no variation in sediment grain size according to vegetation type or coverage at the revegetated site, however, the median diameter of sediment particles at the revegetated site was coarser than at the bare site (fig. 5).

Discussion

29Egashira and Ashida (1981) reported that sediment particles with diameters of less than 0.1 mm were suspended in flowing water in mountainous drainage basins. In this study a grain-size analysis showed that about 90% of the sieved ash used in the treatment had grain sizes of less than 0.1 mm. Suspended sieved ash probably moved together with overland flow. Suspended sediment in flowing water does not accumulate on the stream bed (Major, 2004), and so material that accumulated at the sediment trap would have been predominantly bedload. The overland flow has probably transported the sieved ash to rivers. This is supported by our data, which indicates that the median diameter of trapped sediment particles was larger than 0.1 mm between 21 June and 16 November 1999 (fig. 5). Immediately after the treatment (i.e. between 11 and 21 June) however, the median diameter of the sediment at the sediment trap was smaller than 0.1 mm (fig. 5), indicating that overland flow immediately after the treatment contained little bedload. Immediately after the treatment, the sieved ash had not yet migrated into the porous underlying gravel layer (fig. 7C, G), but was still distributed on the ground surface (fig. 7D, SaG). The overland flow that took place between 11 June and 21 June 1999 clearly moved the sieved ash that had been spread over the ground surface. The silt-sized sieved ash rapidly moved through the subsurface during the rainfall, so that the subsurface soil porosity decreased (fig. 7D, SaG). Hydraulic conductivity of the ground surface deposit experimentally covered with the 2-cm thick sieved ash layer was lower (3.30 x 10-4 cm/s-1) than that of the deposit itself (Ogawa et al., 2006). Both Sa and SaG layers were probably low permeable subsurface layers.

30The rise of Sb/Qb in mid-June 1999 (fig. 5) supports the idea that the sieved ash movement at the subsurface produced the SaG layer during the rainfall, resulting in low infiltration capacity of the ground surface and the increase in overland load at the bare site. The decline in Sb/Qb in September 1999 (fig. 5) provides evidence that the sieved ash in the SaG layer was gradually removed by overland flow, so that the infiltration capacity of the ground surface increased and overland load decreased at the bare site.

31Overland flow was higher after the treatment, and took place under lower rainfall than before the treatment (tab. 3; fig. 6). This indicates that the rainfall required to cause an overland flow event was lower after the sieved ash treatment, and that overland flow occurred more readily after treatment rather than before. The change in rainfall conditions and the increase in overland flow volume after the treatment suggest that a tephra mantle has reduced the infiltration capacity of the ground surface.

32Although sediment load per unit area was lower at the revegetated site than at the bare site, the median diameter of sediment was larger at the revegetated site than at the bare site (fig. 5). A field inspection of the revegetated site showed that there were fewer plants growing inside than near a rill, that herbaceous plants along the rill were toppled by overland flow, and that the sediment inside the rill consisted of coarser gravel clasts than those at the bare site. This suggests that the sediment trapped at the revegetated site was sediment removed from along the rill.

33Post-eruption sediment load per unit area was lower from the revegetated site than from the bare site. Nishida et al. (1998a), who evaluated sediment load from bare and revegetated sites two years after an eruption of Unzen, have obtained similar results. The present results prior to the sieved ash treatment (tab. 4; fig. 4) show that revegetated plants controlled the movement of sediment on sloped surfaces of pyroclastic-flow deposits. The decline in Sr/Qr can be attributed to two phenomena. Firstly, the majority of the sediment was most likely removed by overland flow in 1999, when annual rainfall and overland flow were at their highest. Secondly, steady plant growth at the revegetated site controlled the movement of destabilised sediment. The changes in the S/Q suggest an influence of vegetation cover and volcanic ash mantle on the overland flow and sediment load. At this stage we cannot quantitatively discuss the vegetation capacity to retain the surface soil and to increase infiltration rate of the surface soil; further study is necessary.

34The results of the site monitoring and the vegetation survey support the contention that the decrease in sediment load from slopes overlain by pyroclastic-flow deposits involves two processes. Firstly, volcanic ash during the eruption was remobilised on the ground surface, resulting in increased infiltration capacity of the ground surface and decreased overland flow. Secondly, the growth of plants following the eruption stabilised erodible sediments on the ground surface of the pyroclastic-flow deposits.

Conclusions

35Monitoring at two field sites (bare and revegetated) on Mount Unzen’s slopes overlain by pyroclastic-flow deposits produced three results.

36Firstly, Sr/Qr was greater at the revegetated site than at the bare site prior to the sieved ash treatment, but greater at the bare site than at the revegetated site after the treatment (fig. 5). This shows that veneering the ground surface with sieved ash increased not only overland flow but also sediment load owing to the low infiltration capacity of the ground surface.

37Secondly, changes in Sb/Qb during the sieved ash treatment corresponded well with changes in the accumulation of sieved ash that remained on the ground surface (figs. 5 and 7). This suggests that overland flow from the bare site began to decrease soon after the eruption due to the loss of tephra from the ground surface prior to vegetation re-growth. This is in agreement with the conclusion of an earlier report, which showed that overland flow and surface erosion at the Mount Usu volcano decreased as a result of tephra erosion (Yamamoto and Imagawa, 1983; Yamamoto, 1984).

38Thirdly, this study showed experimentally that changes in the accumulation of sieved ash on the ground correspond with changes in the removal of sieved ash. There are distinct temporal changes in the occurrence of overland flow and sediment load from ground surfaces that are covered by pyroclastic-flow deposits during eruptions. S/Q proved to be an effective index of changes in ground surface conditions. For example, plotting S/Q (fig. 5) clearly shows that both overland flow and sediment load at the bare site were greater after ash spreading. Temporal changes in S/Q reflect changes in the dynamics of sediment load, which depended on the accumulation of sieved ash and plant growth. S/Q can be used only when sediment load is detectable. This highlights the need to devise methods for quantifying the effects of vegetation on overland flow under conditions of low sediment load.

39Data obtained from monitoring experiments in the field show that temporal changes in fine-grained fresh volcanic ash accumulation and vegetation coverage correlate with temporal changes in the ratio of overland flow to sediment load (S/Q). This ratio is, therefore, a useful index of ground-surface stabilisation in pyroclastic-flow deposits.

Haut de page

Bibliographie

Adikari Y., Kikuchi S., Araya T. (2005) – Sedimentation and succession pattern in regenerating primary vegetation governing gully formation in the area devastated by the 2000 eruption of the Mount Usu volcano, in Hokkaido, Japan. Zeitschrift für Geomorphologie, Supplement band, 140, 103-113.

Collins B. D., Dunne T. (1986) – Erosion of tephra from the 1980 eruption of Mount St. Helens. Geological Society of America Bulletin 97, 896-905.

Egashira S., Ashida K. (1981) – The productive areas and transporting process of wash load in mountainous drainage basins. Annuals of Disaster Prevention Research Institute Kyoto University, 24B-2, 239-250 (in Japanese, with English abstract).

Harada T., Hirano M., Kawahara K. (1999) Runoff analysis of debris flow in the Mizunashi River at Mt. Unzen. Journal of the Japan Society of Erosion Control Engineering 52, 1, 3-9 (in Japanese, with English abstract).

Ishikawa Y., Yamada T., Chiba T. (1996) – Topographic changes on the Unzen volcano due to volcanic activities from 1991 to 1995. Journal of the Japan Society of Erosion Control Engineering 49, 1, 38-44 (in Japanese, with English abstract).

Jitousono, T., Shimokawa E., Tsuchiya S. (1996) – Debris flow following the 1984 eruption with pyroclastic flows in Merapi volcano, Indonesia. Journal of the Japan Society of Erosion Control Engineering, Special Issue, 48, 109-116.

Major J. J. (2004) – Post-eruption suspended sediment transport at Mount St. Helens: Decadal-scale relationships with landscape adjustments and river discharges. Journal of Geophysical Research 109, F01002, doi:10.1029 /2002JF000010, 1-22.

Miyabuchi Y. (1999) – Deposits associated with the 1990-1995 eruption of Unzen volcano, Japan. Journal of Volcanology Geothermal Research 89, 139-158.

Nagasaki prefectural government, Shimabara branch office (2000)First step of revegetation in Mount Unzen-Fugendake. 292 p. (in Japanese)

Nishida K., Kobashi S., Mizuyama T. (1996) – Change of infiltration rate of the deposit in Mount Unzen-Fugendake. Journal of the Japan Society of Erosion Control Engineering 49, 1, 49-53 (in Japanese).

Nishida K., Kobashi S., Mizuyama T. (1998a) – How has revegetation changed surface runoff and sediment discharge on pyroclastic sediments on Unzen volcano? Journal of the Japanese Society of Revegetation Technology 23, 4, 249-255 (in Japanese, with English abstract).

Nishida K., Mizuyama T. (1998b) – Surface structure of pyroclastic deposits on Unzen volcano. Journal of the Japan Society of Erosion Control Engineering 51, 3, 44-46 (in Japanese).

Ogawa Y., Shimizu A., Kubotera H. (2006) – The influence of soil micromorphology on surface soil permeability of slopes overlaid with new volcanic deposit of Mount Fugen. Journal of the Japanese Forest Society 88, 5, 329-336 (in Japanese, with English abstract).

Pierson T. C., Janda, R. J. (1992) – Immediate and long-term hazards from lahars and excess sedimentation in rivers draining Mt. Pinatubo, Philippines. U.S. Geological Survey Water-Resources Investigations Report, 92-4039, 1-35.

Yamakoshi T., Suwa H. (1998) – Responses of surface runoff and sediment discharge to revegetation on the slope of pyroclastic-flow deposits, Mount Unzen, Japan. Journal of the Japan Society of Erosion Control Engineering 51, 3, 3-10 (in Japanese, with English abstract).

Yamamoto H., Imagawa T. (1983) – Surface runoff on a slope covered by 1977-1978 tephra from Usu volcano. Hydrology 13, 25-33 (in Japanese, with English abstract).

Yamamoto H. (1984) – Erosion of the 1977-1978 tephra layers on a slope of Usu volcano, Hokkaido. Transactions Japanese Geomorphological Union 5, 2, 111-124 (in Japanese, English abstract).

Haut de page

Annexe

Version française abrégée

Le volcan Unzen à l’ouest de l’île de Kyushu, au sud-ouest du Japon, est entré en éruption en novembre 1990 (fig. 1). Plus de 9 400 écoulements pyroclastiques ont été produits de 1990 à 1995 essentiellement sous l’effet des écroulements liés à la croissance du dôme. Autour de celui-ci ont été déposées d’épaisses couches de matériaux volcaniques issus des épisodes de forte activité éruptive (fig. 2) et du remaniement concomitant des particules meubles, en particulier par des coulées de débris. La quantité de sédiments mobilisés sur les pentes et le nombre de coulées de débris diminuèrent rapidement dès que l’activité éruptive cessa (Yamamoto, 1984 ; Collins and Dunne, 1986 ; Pierson and Janda, 1992 ; Jitousono et al., 1996).

Le but de cette étude est de mettre en évidence les facteurs qui contribuent à faire diminuer la charge solide entraînée par le ruissellement après les éruptions. Sur des parcelles expérimentales, le ruissellement, la charge solide et la couverture végétale ont fait l’objet de mesures et de relevés sur une pente artificiellement ensemencée, ainsi que sur une pente au sol nu, mais saupoudrée, en juin 1999, par un voile de cendre volcanique finement tamisée (tab. 2 ; fig. 3). Le site végétalisé a été observé d’août 1998 à décembre 2001 tandis que le site dépourvu de végétation a été observé d’août 1998 à novembre 1999.

Les campagnes de mesures et d’observations apportent les éclaircissements suivants : sur le site au sol nu, immédiatement après l’apport de cendre tamisée, le débit de pointe maximum du ruissellement fut 1,4 fois supérieur et le ruissellement total fut deux fois plus important qu‘auparavant et que sur le site végétalisé pour une même magnitude des abats d’eau (tab. 3 ; fig. 6). Sur le site revégétalisé, les eaux ruisselées annuellement et l’entraînement de particules ont graduellement diminué au fur et à mesure de la croissance de la végétation, si bien que le volume de sédiments déplacés n’était plus détectable dès 2001 (tab. 4). Le quotient entre la charge solide transportée par le ruissellement et le volume total ruisselé (soit S/Q) est considéré comme un indice qui varie selon le degré de stabilisation de la surface des dépôts pyroclastiques. Les modifications de ce quotient S/Q (fig. 5, fig. 7 et fig. 8) sont corrélées avec l’inégale mobilisation dans le temps des cendres fines fraîchement retombées et la colonisation par la couverture végétale.

Les résultats suggèrent deux explications qui rendent compte de la diminution de la charge solide entraînée par le ruissellement sur les pentes inscrites dans des dépôts pyroclastiques. La première s’appuie sur le fait que la cendre volcanique fine, retombée durant les éruptions, joue le rôle d’écran imperméable ; le décapage rapide des particules aisément mobilisables rétablit la capacité d’infiltration du sol avec comme corollaire une diminution du ruissellement. La seconde est due à la croissance des plantes qui, colonisant les pentes après l’éruption, stabilisent les matériaux meubles à la surface des dépôts pyroclastiques.

Haut de page

Table des illustrations

Titre Fig. 1 – Location of field sites.Fig. 1 – Emplacement des sites d’étude.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-1.jpg
Fichier image/jpeg, 236k
Titre Fig. 2 – Oblique aerial view of part of the Unzen volcano in 1997, from the southeast. Photograph was taken by the Kyushu Regional Forest Office.Fig. 2 – Depuis le sud-est, vue aérienne oblique d’une partie du volcan Unzen en 1997. Photo prise par le Bureau Forestier Régional de Kyushu.
Légende BS: bare site; RS: revegetated site.BS : site dénudé ; RS : site revégétalisé.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-2.jpg
Fichier image/jpeg, 804k
Titre Fig. 3 – Topographic map and equipment at the bare site.Fig. 3 – Carte topographique et équipement du site dépourvu de végétation.
Légende 1: rain gauge; 2: weir and sediment trap; 3: sieved ash treatment area; 4: plate dividing watershed; 5: gully dividing watershed; 6: natural topographic boundary to the watershed; 7: water flow in a heavy rainfall event; 8: surface soil monitoring area.1 : pluviomètre ; 2 : déversoir et piège à sédiments ; 3 : aire saupoudrée de cendres tamisées ; 4 : tôle délimitant l’aire contributive du ruissellement ; 5 : ravine délimitant une aire contributive du ruissellement ; 6 : limite naturelle de bassin versant ; 7 : cheminement de l’eau en cas de fortes pluies ; 8 : aire de la parcelle expérimentale.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-3.jpg
Fichier image/jpeg, 260k
Titre Fig. 4 – Hydrograph of overland flow at the field sites. Hydrograph of the bare site before the sieved ash treatment.Fig. 4 – Hydrogramme du ruissellement sur les sites d’étude. L’hydrogramme du site dénudé a trait à la période précédant l’apport de cendres tamisées.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-4.jpg
Fichier image/jpeg, 108k
Titre Fig. 5 – Temporal changes in total rainfall (R), overland flow (Qb and Qr), sediment load, over-land flow ratio (Sb/Qb and Sr/Qr) and median sediment clast diameter (d50) collected in the sediment traps at the two field sites.Fig. 5 – Variabilité des précipitations totales (R), du ruissellement (Qb et Qr), de la charge solide, du rapport entre charge solide et ruissellement (Sb/Qb et Sr/Qr) et de la médiane du diamètre (d50) des particules récoltées dans les pièges à sédiments sur les deux sites d’étude.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-5.jpg
Fichier image/jpeg, 352k
Titre Fig. 6 – Hydrograph of overland flow at bare site before (A) and after (B) the sieved ash treatment.Fig. 6 – Hydrogramme du ruissellement sur la parcelle nue avant (A) et après (B) le saupoudrage de cendres tamisées.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-6.jpg
Fichier image/jpeg, 184k
Titre Fig. 7 – View and surficial sediment profile of the sieved ash treatment area. A: view before the ash treatment; B: view immediately after the ash treatment; C: soil profile immediately after the ash treatment; D: soil profile one month after the ash treatment; E: view one month after the ash treatment.Fig. 7 – Vue et coupes des sédiments superficiels de la parcelle ayant été saupoudrée de cendres tamisées. A : vue avant l’apport des cendres ; B : vue immédiatement après le saupoudrage des cendres ; C : profil du sol immédiatement après l’apport des cendres ; D : profil du sol un mois après le saupoudrage des cendres ; E : vue un mois après l’apport des cendres.
Légende Pr: surface profile monitoring area; Sa: sieved ash; G: gravel layer; SaG: sieved ash within gravel layer; P: pyroclastic-flow deposit beneath gravel layer.Pr : parcelle étudiée ; Sa : cendres tamisées ; G : couche de graviers ; SaG : cendres tamisées infiltrées dans la couche de gravier ; P : dépôt de coulée pyroclastique sous la couche de graviers.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-7.jpg
Fichier image/jpeg, 752k
Titre Fig. 8 – Changes in mean growth height, coverage ratio and vegetation type at the revegetated site. A: view in July, 1998; B: view in August, 2001. The 1996-1999 data are from the Shimabara branch office of Nagasaki prefectural government (2000) and the 2000-2001 data is from the authors.Fig. 8 – Accroissement de la taille moyenne des végétaux, du taux de couverture et du type de végétation sur le site revégétalisé. A : vue en juillet 1998 ; B : vue en août 2001. Les données de 1996 à 1999 proviennent du bureau de Shimabara de la préfecture de Nagasaki (2000), tandis que celles de 2000 à 2001 proviennent des auteurs.
URL http://geomorphologie.revues.org/docannexe/image/3962/img-8.jpg
Fichier image/jpeg, 695k
Haut de page

Pour citer cet article

Référence papier

Yasuhiro Ogawa, Hiromu Daimaru et Akira Shimizu, « Experimental study of post-eruption overland flow and sediment load from slopes overlain by pyroclastic-flow deposits, Unzen volcano, Japan », Géomorphologie : relief, processus, environnement, vol. 13 - n° 3 | 2007, 237-246.

Référence électronique

Yasuhiro Ogawa, Hiromu Daimaru et Akira Shimizu, « Experimental study of post-eruption overland flow and sediment load from slopes overlain by pyroclastic-flow deposits, Unzen volcano, Japan », Géomorphologie : relief, processus, environnement [En ligne], vol. 13 - n° 3 | 2007, mis en ligne le 01 octobre 2009, consulté le 18 octobre 2017. URL : http://geomorphologie.revues.org/3962 ; DOI : 10.4000/geomorphologie.3962

Haut de page

Droits d’auteur

© Groupe français de géomorphologie

Haut de page
  • Logo CNRS - Institut des sciences humaines et sociales
  • Logo Groupe français de géomorphologie
  • Revues.org