Réf. Foster & al. 2003 - A

Référence bibliographique complète

FOSTER, G.C., DEARING, J.A., JONES, R.T., CROOK, D.S., SIDDLE, D.J., HARVEY, A.M., JAMES, P.A., APPLEBY, P.G., THOMPSON, R., NICHOLSON, J., LOIZEAU, J.-L. 2003. Meteorological and land use controls on past and present hydro-geomorphic processes in the pre-alpine environment: an integrated lake–catchment study at the Petit Lac d’Annecy, France. Hydrological Processes, Vol. 17, 3287-3305.

Abstract: A wide range of environmental records is integrated in order to reconstruct the mechanisms of flooding and sediment transport within the 170 km² Petit Lac catchment, Annecy, France, over time scales of 10-1 to 102 years. These records include sequential lake sediment trap samples and cores, floodplain stratigraphies, dated landform assemblages, hydro-meteorological records, and documented histories of river channel and land-use change. Mineral magnetic measurements are used as the basis for classifying catchment sediment sources and tracing sediment movements through time. Records of magnetic susceptibility for monthly sediment trap samples (1998–99) track seasonal discharge, peaking in winter and spring. Magnetic records in lake sediment cores are compared against and tuned to precipitation records to provide dated proxy records for past discharge spanning sub-annual to decadal time scales back to 1826. Calculated sediment accumulation rates in lake sediment cores are used as proxies for time-averaged catchment sediment load. Analysis of the results reveals that climate and land-use controls on the hydrological and sediment system are complex and vary according to the time scale of observation. In general, cycles of agricultural expansion and deforestation appear to have been the major cause of shifts in the sediment system through the late Holocene. Deforestation in the 18th century may have caused a number of high-magnitude flood and erosion events. As the time scale of observation becomes shorter, changes in climate and hydro-meteorological conditions become progressively more important. Since the mid-19th century, smoothed records of discharge roughly follow annual precipitation; this is in contrast to sediment load, which follows the trend of declining land-use pressures. Episodic erosion events during this recent period seem to be linked to geomorphic evidence for slope instability in the montane and sub-alpine zones, triggered by intense summer rainfall. At the annual scale, changes in seasonal rainfall become paramount in determining sediment movement to downstream locations. The study demonstrates that the connections between forcings and responses span a four-dimensional array of temporal and spatial scales, with strong evidence for dominantly nonlinear forcing–response mechanisms.

Mots-clés
Lac d’Annecy; lake sediments; mineral magnetism; erosion; flooding; human impact; climate

Organismes / Contact

• Department of Geography, University of Liverpool, Roxby Building, Liverpool L69 7ZT, UK 2 - Correspondance: gfoster@liv.ac.uk
• Department of Applied Mathematics, University of Liverpool, Liverpool, UK 3
• Department of Geology and Geophysics, University of Edinburgh, Kings Building, Edinburgh, UK 4
• Centre d’études en Sciences Naturelles de l’Environnement, Institut F.-A. Forel, Université de Genève, 1290 Versoix, Switzerland


(1) - Paramètre(s) atmosphérique(s) modifié(s)
(2) - Elément(s) du milieu impacté(s)
(3) - Type(s) d'aléa impacté(s)
(3) - Sous-type(s) d'aléa
Precipitation Soil erosion Floods  

Pays / Zone
Massif / Secteur
Site(s) d'étude
Exposition
Altitude
Période(s) d'observation
France Haute-Savoie (Alpes du Nord) Lac d’Annecy catchment and sub-catchments of three major streams flowing into the Petit Lac: the Eau Morte, Ire and Bornette   400-2200m •1975-1999
•1825-1995
•Discussion over millennial time scales

(1) - Modifications des paramètres atmosphériques
Reconstitutions
 
Observations

 

Modélisations

A review of projected climatic change over the next century in the western French Alps anticipates increased mean temperatures and precipitation (Gyalistras et al., 1998).

Hypothèses
 

Informations complémentaires (données utilisées, méthode, scénarios, etc.)
 

(2) - Effets du changement climatique sur le milieu naturel
Reconstitutions

Channel changes, historical flooding and floodplain development:
The most dramatic channel changes on the lower floodplain since before 1730 have included at least four shifts of the Ire torrent. The position of the 18th century Ire is marked by linear channel fills on the floodplain surface that appear to reflect natural avulsions prior to 1730. Written accounts of the flooding of Verthier would suggest that the 1906 course was established before 1825 and remained until around 1918, after which the modern dyke was constructed. Historical records spanning 1725–45 suggest that the frequency of flooding of the main channels peaked on five occasions, ca 1750, 1800, 1850, 1875, and 1900 (Crook et al., 2002). [...]

The presence of a consistent suite of coarse horizons and palaeochannels along the length of the upper floodplain provides strong evidence for recent flooding and river channel changes. The channels contain imbricated cobbles set within sandy gravels, and indicate a final stage of channel adjustment and rapid floodplain accretion before major dyking operations in the early 20th century. Summary results of channel studies show that a reduction of width : depth ratios between the lower and middle stages was accompanied by a shift to assymetrical channel geometry, indicative of meandering. This signals a reduction in gradients, perhaps in response to reduced sediment calibre, sediment load or stream power. The upper channel signals a return to a more unstable system, in which braiding occurred possibly in response to increased sediment supply. The upper channel contains anthropogenic artefacts (e.g. brick fragments and a bed spring), linking it to the channel depicted on the 1906 cadastral map. Two radiocarbon measurements of macrofossils and charcoal in a flood horizon ~30 cm below the base of the channels give modern dates with maximal ages of ~280 cal. years BP. Therefore, the channel fills reflect rapid aggradation within the recent historical period, perhaps post-1700. Two radiocarbon measurements of wood fragments in organic-rich gravels at the contact with underlying clays and overlying silts give dates of 2351–2746 and 2212–2713 cal. years BP respectively, suggesting that overbank deposition started during Iron Age times and that no major coarse flood deposits were laid down between ~2500 and ~300 BP.

[See also Sub-alpine slope instability, p. 3293]

Recent lake sediments: sub-annual process-response [1973-1978]:
From 1975 (the date of the oldest discharge records), peak values of frequency-dependent magnetic susceptibility show a tendency to rise with increasing flood magnitude, suggesting that the entrainment of lowland soil (free-draining soils on lower slopes and flatter areas of montane valleys) takes place progressively as lowland valleys flood and as sediment in temporary channel storage sites becomes mobilized. Declining sediment accumulation rates during the late 1970s coincide with a major change in the seasonal distribution of precipitation, from summer/autumn peaks during 1973–77 to a winter/spring maximum in 1978, suggesting that sediment accumulation rates are controlled by the seasonality of rainfall, and particularly the occurrence of highly erosive, intensive rainstorms.

Lake sediments: the past two centuries [1826-1995]:
The successful calibration of magnetic data with hydro-meteorological data over seasonal and annual time scales is extended to the past two centuries in the upper 1 m of a sediment sequence collected from the central plain of the Petit Lac (water depth 55 m). The main features in the 5 year smoothed record of precipitation at Geneva (~50 km north of Annecy) extending back to 1826 correspond reasonably well with the low-frequency magnetic susceptibility record in the sediment sequence, suggesting that low-frequency magnetic susceptibility may also be used as a first-order proxy of discharge and precipitation over longer time scales. The accuracy and precision of the tuned chronology for this central core is confirmed by 210Pb and 137Cs measurements and a pollen-based event chronology (Jones et al., in preparation), and is used to calculate mean sediment accumulation rates. The trend in the discharge proxy record shows gradually rising peak and minimum values from 1826 to the present. In contrast, the sediment yield record shows a declining trend of minimum values from ~0.63 to 0.1 g cm-2 year-1 with five or six periods of high sediment accumulation rates ranging from ~1.25–0.6 g cm-2 year-1. There is only a partial correspondence between the peak discharge and peak sedimentation rate, suggesting that annual river discharge records are more closely related to annual meteorological conditions, and annual precipitation in particular. Therefore, it appears that sediment supply to the lake over the longer time scale is controlled to a large extent by factors other than annual and seasonal meteorology.

Discussion: Hydrological and sediment dynamics:
Over millennial time scales, long sediment records in the Petit Lac may be compared with pollen diagrams, regional land-use syntheses and climate reconstructions to demonstrate that shifts in the sediment supply have been forced largely by land-use changes, especially upland deforestation in the period 800–1100 AD (Dearing et al., 2001). Decadal, annual and sub-annual proxy records, presented here for discharge and sediment load within the past 200 years, offer the opportunity to extend the analysis of climate or land use as forcing factors at different time scales. For recent centuries, documentary and census data for the catchment provide information about potential land-use forcings on the hydrological cycle in the catchment [...] Comparison of these records with the proxy discharge record shows that ‘higher than expected’ peaks in the low-frequency magnetic susceptibility record ca 1890 and during the 1960s might argue for amplified flood levels caused by pre-1900 deforestation and 20th century dyking, urbanization and intensive cultivation. But otherwise, annual flood magnitudes appear to be highly tuned to annual precipitation: large changes in land use have not triggered systematic shifts in rainfall–runoff relationships. The timing of widespread floodplain instability and destructive flooding within the past 300 years is not sufficiently constrained but is most likely to pre-date the mid 19th century. Phases of deforestation along the main river valleys were partly driven by developing proto-industrial activities from the 17th century onwards, reaching levels as high as 10% of all forest per year in the late 18th and early 19th centuries. This suggests that the important zone of runoff generation, at high montane and sub-alpine altitudes above the tree-line (~1500 m) where precipitation levels are highest, has not been affected by land-use transformations over the past few centuries, though this is unlikely to have remained the case over millennial time scales. In contrast, declining sediment load since 1825, a trend that is also observed at a finer time scale since 1974, suggests that the reductions in agricultural intensity and improved forest conservation have affected baseline sediment supplies. Short-term peaks in sediment load since 1825 coincide with short periods of deforestation before 1850, ca 1880 and post-war cultivation practices 1945–70, but the timing of sediment pulses since ca 1880 is also closely linked to the widespread activation of debris flows every 20–40 years. The evidence suggests that, although underlying trends of sediment supply are driven by land use, this does not exclude short-term sediment responses to changes in climatic seasonality, or seasonal to inter-annual changes of land use and agricultural practices driven by technological or socio-economic factors. At the decadal time scale, climate and land use may interact on the sediment system in complex ways that are difficult to resolve.

At the annual time scale, sediment load is at a maximum at the time of snowmelt, and in recent decades it is clear that the magnitude is greatly amplified by the erosive effects of rainfall in the preceding summer/autumn months. This lag-time represents the period between activation of largely upland slope erosion processes delivering unweathered substrate and the onset of efficient slope–channel coupling and sediment transport mechanisms during snowmelt. Additional surface soil is incorporated into the sediment load during snowmelt from sources close to montane and lowland channels, presumably by overbank flooding. Less common are discrete sediment pulses, usually dominated by topsoil sources, transported during intense storms outside the snowmelt period. Winter snowfall and rainfall are therefore second-order factors in controlling the magnitude of the sediment pulse. Summer storms may be effective in triggering rill processes, but they are only rarely effective at delivering sediment from surfaces to channels. However, their importance in determining sediment supplies at all time scales longer than the season should not be understated. Intense summer rainfall probably gives rise to similar frequency–magnitude relationships for two independent sets of processes within short and medium time scales: upland mass movements and lowland soil erosion.

The effects of climatic and land-use forcings on discharge and sediment load are clearly related to the time scale of observation. In general, cycles of agricultural expansion and deforestation appear to have been the major cause of shifts in the sediment system through the late-Holocene. As the time scale of observation becomes shorter, changes in climate and hydro-meteorological conditions become progressively more important. At the annual scale, changes in seasonal rainfall become paramount in determining sediment movement to downstream locations. We may certainly hypothesize that land-use changes have conditioned and, in the present situation, amplified later responses in the sediment system. This particularly applies to the early deforestation of the montane zone and cultivation in Iron Age and early medieval times (Dearing et al., 2001) that caused the rapid aggradation of overbank deposits on the valley floor after ~2500 cal. years BP. It also applies in the 18th and 19th centuries, when dramatic phases of deforestation led to rapid increases in sediment supply and floodplain accretion, channel changes and destructive flooding. These events may have triggered the transgression of sediment thresholds on steep slopes, which in turn increased their sensitivity to mass movement failure during extreme rainfall events in the 19th and 20th centuries.

Observations
 
Modélisations
 
Hypothèses

In a warmer and wetter climate, we might expect mean annual flood levels to follow roughly the late winter and spring precipitation. But careful management of the montane and sub-alpine zones seems critical for avoiding flashy hydrological responses, especially in view of the importance of erosive summer rainfall in determining to a large extent the annual sediment supply.


Sensibilité du milieu à des paramètres climatiques
Informations complémentaires (données utilisées, méthode, scénarios, etc.)

Contemporary sediment traps:
Comparison of selected magnetic and physical properties of sediments for each sampling period with summed monthly and daily discharge shows that bulk low-frequency magnetic susceptibility varies cyclically from minima in summer to a maximum in winter and spring. [...] The seasonal cyclicity of magnetism in the trap samples indicates that low-frequency magnetic susceptibility may provide a sensitive proxy for mean monthly stream discharge (QM) that could be applied to lake sediment cores. [See details in the study]

Recent lake sediments: inter-annual process-response:
[...] Monitored records of the monthly mean of daily discharge QM for the Eau Morte are available for part of this time scale, 1975–99, corresponding to the upper ~55 cm of the core, and records of monthly precipitation PM at Annecy are available for most of the period back to 1873. To test the idea that low-frequency magnetic susceptibility may be used as a discharge proxy, as described above for the sediment traps, the QM and PM records were smoothed using a 12-point moving average to account for the time-integrated effects of mixing and bioturbation within the sediment record. The resulting filtered time series for discharge (Q12) and precipitation (P12) were compared with the sediment low-frequency magnetic susceptibility record. [...] A comparison between the dated low-frequency magnetic susceptibility record and precipitation data for the period covering the section of sediment shows a good correspondence with P12 and an even stronger match to a six-point smoothed curve (P6). [...] The findings of this study provide strong evidence for hydrological forcing of the recent sediment low-frequency magnetic susceptibility signal on annual to sub-annual time scales and confirm the use of low-frequency magnetic susceptibility as a sedimentary proxy for monthly discharge and precipitation.

Using the hydro-meteorologically tuned chronology for a marginal core (1 m long) taken from 28 m water depth ~300 m north of the main Eau Morte delta, sediment accumulation rates are calculated for 29 time periods between 1973 and 1999. The mean sediment accumulation rate is 3 g cm-2 year-1 with a declining trend towards the present and episodes of high sediment accumulation rates reaching 16 g cm-2 year-1 between 1973 and 1978, and during 1989. The maximum rate appears to coincide with peak precipitation levels in 1977, but high rates over the period 1973–76 and in 1989 correspond with periods of relatively low precipitation. This suggests that runoff is not always the dominant control on catchment sediment loads. Rather, peak sediment load is more frequently controlled by sediment supply and availability than by precipitation and discharge.

Recent lake sediments: sub-annual process-response:
During the period 1973–78 the presence of covarying high values of sediment accumulation rate, frequency-dependent magnetic susceptibility and proportions of paramagnetic material, especially in the winter–spring seasons of 1973, 1975, 1976–77 and 1978, emphasize the importance of snowmelt processes for the coupling of upland and lowland sediment sources with the fluvial network. During this period, the covarying trends of sediment accumulation rate and proportions of paramagnetic material suggest that upland-derived material (primary minerals derived from high montane/sub-alpine soils and parent material) provides the dominant sediment source and exerts the strongest control on sediment loads. From 1975 (the date of the oldest discharge records), peak values of frequency-dependent magnetic susceptibility show a tendency to rise with increasing flood magnitude, suggesting that the entrainment of lowland soil (free-draining soils on lower slopes and flatter areas of montane valleys) takes place progressively as lowland valleys flood and as sediment in temporary channel storage sites becomes mobilized.

Declining sediment accumulation rates during the late 1970s coincide with a major change in the seasonal distribution of precipitation, from summer/autumn peaks during 1973–77 to a winter/spring maximum in 1978, suggesting that sediment accumulation rates are controlled by the seasonality of rainfall, and particularly the occurrence of highly erosive, intensive rainstorms. Furthermore, daily discharge data reveal that each of the years 1975, 1976 and 1977 witnessed one high-magnitude (mean Q > 104 l s-1) summer/autumn storm event. Therefore, relatively high sediment contributions from both upland and lowland sources, and the high sediment accumulation rate values, reflect the supply and availability of sediment following intensive rainstorms—often several months prior to deposition in the lake. The seasonal lag between peak precipitation in summer/autumn and the winter/spring sediment delivery by snowmelt presumably represents the temporary storage of eroded material at slope base and near-channel locations. The magnitude of snowmelt is far less important in determining sediment loads than the degree of slope–channel coupling and the size of potential sediment stores.

The authors reassess the significance of lake sediment magnetism proxies for reconstructing records of soil erosion over 10-1 to 102 year time scales in the light of new information from sediment trap sampling, marginal lake sediments, field geomorphological studies, and hydro-meteorological data, supplemented with documentary records of flooding, land use and channel changes. The aims of this work are: (1) to establish calibrations between contemporary and recent magnetic signals in trapped and lake inflow sediments; (2) to assess the extent to which geomorphic proxies can be extrapolated to the long-term record; and (3) to make a first attempt at explaining the significance of hydro-meteorological and land-use forcings on flooding and erosion over the past few centuries.

A key method employed in earlier studies, and this, has been the use of mineral magnetic properties (Thompson and Oldfield, 1986; Walden et al., 1999) to define sediment sources and to infer erosional processes in sediment records. Limestones that are very poor in primary magnetic minerals dominate the geological units in the catchment. The non-diamagnetic properties of the limestones are almost completely dominated by goethite, and the marls contain iron-bearing paramagnetic minerals and goethite (Hu, 1997). Rissian and Würmian glaciers (Evin et al., 1994) deposited at least two types of till: one contains abundant schist erratics derived from the inner Alps and is dominated by paramagnetic minerals with low concentrations of primary ferrimagnetic minerals (Hu, 1997); the other is composed of limestones and marls considered to be of local origin (Beck et al., 1996; Hu, 1997; Hu et al., 2001). Early magnetic and Mössbauer studies showed that many of the surface soils in the well-drained parts of the catchment are relatively rich in fine-grained secondary ferrimagnetic minerals (SFMs) produced by pedogenic enhancement mechanisms (Dearing, 1979; Longworth et al., 1979; Dearing et al., 1996). The main factors that control SFM concentrations appear to be drainage and microclimate, with lithology exerting a major local influence only where glacial tills contain primary magnetite and possibly haematite in schist erratics. Soil magnetic maps (Dearing et al., 2001) for the Petit Lac catchment show four main groups of soil, used in the present study as the basis for sediment source ascription.

Materials and Methods:
Lake sediment traps and cores: Two sediment traps were suspended in the connected Grand Lac at water depths of 20 and 46 m. The traps were emptied on 18 visits between April 1998 and December 1999, with trap periods ranging between 20 and 41 days. The samples collected were dried for a range of analyses, which included mineral magnetism [...]. [After treatment], magnetic analyses included standard measurements (Thompson and Oldfield, 1986; Walden et al., 1999) of low- and high-frequency magnetic susceptibility (permitting the calculation of frequency-dependent magnetic susceptibility), susceptibility of anhysteretic remanence, isothermal remanence and in-field magnetization measured in a vibrating sample magnetometer. Full details are presented in Dearing et al. (2001).

Geomorphological studies: Floodplain surveys were carried out in autumn 1999 in order to identify the nature and sequence of fluvial deposition on the Eau Morte valley floor [...], by gouge coring. At some sites, the uppermost channel fills were morphologically and sedimentologically described in order to determine the nature of channel changes prior to modern dyking operations. Lichen-based slope geomorphic studies were carried out in autumn 1998 in order to investigate the timing of regional sub-alpine debris-flows (above ~1500 m) in the Petit Lac catchment and the Chaine des Aravis, some ~10 km to the east of the lake. [...] This study focuses on the findings from two debris-flow sites, at the Alp de Fier and Col de l’Arpettaz, representative of slope processes operating between 1500 and 2100 m on hillslope and mountainfront landscape units respectively.


(3) - Effets du changement climatique sur l'aléa
Reconstitutions
 
Observations
 
Modélisations
 
Hypothèses

In a warmer and wetter climate, we might expect mean annual flood levels to follow roughly the late winter and spring precipitation. But careful management of the montane and sub-alpine zones seems critical for avoiding flashy hydrological responses, especially in view of the importance of erosive summer rainfall in determining to a large extent the annual sediment supply.


Paramètre de l'aléa
Sensibilité des paramètres de l'aléa à des paramètres climatiques
Informations complémentaires (données utilisées, méthode, scénarios, etc.)
 

[See above]


(4) - Remarques générales

Introduction:
[...] Previous research at Annecy has demonstrated the use of lake sediment magnetism in reconstructing records of soil erosion (Higgitt et al., 1991) and flooding (Thorndycraft et al., 1998). Using an extended lake–catchment approach, Dearing et al. (2001) proposed that the lake sediment signal could be disaggregated in terms of shifting contributions of sediment derived from magnetically distinct catchment sources that correspond broadly with altitudinal soil zonations. Following a sharp erosional response to forest clearance at ~1000 cal. years BP, the role of low–mid-altitude surface soils and high montane soils as sediment sources show divergent trends (Dearing et al., 2001), with the contribution from the latter gradually increasing up to the present day. This may simply reflect the enhanced storage of surface soil on the floodplain after 2000–1000 cal. years BP. Alternatively, it may imply that while the low–mid-altitude soil–vegetation systems showed some degree of stabilization over subsequent centuries, the high montane zone progressively destabilized.


(5) - Syntèses et préconisations

The present findings are beginning to clarify the connections between forcings and responses across a four-dimensional array of temporal and spatial scales. Not only are the controls on discharge and sediment delivery a consequence of time scale, it is also apparent that soil–vegetation–hydrological systems in diverse altitudinal zones, within the same catchment, have significantly different degrees of resilience to combinations of climate and human activities. The mix of forcings and scales means that some hydro-geomorphic responses are clearly direct, broadly linear and exhibit negligible time lags; other less obvious forcing–response relationships involve long-term and threshold-dependent nonlinear changes. Synergistic interactions between climate, human activities and hydro-geomorphic response in the Annecy catchment are therefore complex. An improved understanding of the relationships between environmental forcings and responses in this and other landscapes requires the adoption of the methodological framework used in this study, in which all relevant spatial and temporal scales are included and integrated.

Références citées :

Beck C, Manalt F, Chapron E, Vanrensbergen P, Debatist M. 1996. Enhanced seismicity in the early postglacial period evidence from the post Würm sediments of Lake Annecy, northern Alps. Journal of Geodynamics 22: 155–171.

Crook DS, Siddle DJ, Jones RT, Dearing JA, Foster GC, Thompson R. 2002. Forestry and flooding in the Annecy Petit Lac catchment, Haute-Savoie, 1730–2000. Environment and History 8: 403–428.

Dearing JA. 1979. The application of magnetic measurements to studies of particulate flux in lake-watershed ecosystems. PhD thesis, University of Liverpool.

Dearing JA. 1999. Holocene environmental change from magnetic proxies in lake sediments. In Quaternary Climates, Environments and Magnetism, Maher BA, Thompson R (eds). Cambridge University Press: Cambridge; 231–278.

Dearing JA, Hay K, Baban S, Huddleston AS, Wellington EMH, Loveland PJ. 1996. Magnetic susceptibility of topsoils: a test of conflicting theories using a national database. Geophysics Journal International 127: 728–734.

Dearing JA, Hu Y, Doody P, James PA, Brauer A. 2001. Preliminary reconstruction of sediment–source linkages for the past 6000 yr at the Petit Lac d’Annecy, France, based on mineral magnetic data. Journal of Paleolimnology 25: 245–258.

Evin J, Bintz P, Monjuvent G. 1994. Human settlements and the last deglaciation in the French Alps. Radiocarbon 36: 345–357.

Foster GC. 2001. The geomorphic significance of lake sediments: an integrated lake–catchment study of process and response in a sub-alpine landscape. Unpublished PhD thesis, University of Liverpool.

Gyalistras D, Schär C, Davies HC, Waner H. 1998. Future Alpine climate. In Views from the Alps: Regional Perspectives on Climate Change, Cebon P, Dahinden U, Davies HC, Imboden DM, Jaeger CC (eds). The MIT Press: Cambridge, MA; 171–223.

Higgitt SR. 1985. A palaeoecological study of recent environmental change in the drainage basin of the Lac d’Annecy (France). PhD thesis, University of Liverpool.

Higgitt SR, Oldfield F, Appleby PG. 1991. The record of land use change and soil erosion in the late Holocene sediments of the Petit Lac d’Annecy, eastern France. The Holocene 1: 14–18.

Hu Y. 1997. A magnetic approach to the establishment of sediment–source linkages for reconstructing the Late Pleistocene and Holocene environmental evolution of the Lac d’Annecy, France. PhD thesis, University of Liverpool.

Hu Y, Oldfield F, Manalt F, Beck C. 2001. The environmental significance of magnetic measurements of Late Pleistocene and Holocene sediment sequence from the Grand Lac d’Annecy, eastern France. Journal of Paleolimnology 25: 193–203.

Jones RT, Foster GC, Dearing JA, Crook DS, Siddle DJ, Appleby PG. In preparation. Reconstructing recent sedimentation patterns in the Petit Lac d’Annecy.

Longworth G, Becker LW, Thompson R, Oldfield F, Dearing JA, Rummery TA. 1979. Mössbauer effect and magnetic studies of secondary iron oxides in soils. Journal of Soil Science 30: 93–110.

Thompson R, Oldfield F. 1986. Environmental Magnetism. George Allen & Unwin: London.

Thorndycraft V, Oldfield F, Hutchinson S, Crooks PRJ, Appleby PG. 1998. A high resolution flood event stratigraphy in the sediments of the Petit Lac d’Annecy. The Holocene. 8: 741–746.

Walden J, Oldfield F, Smith JP (eds). 1999. Environmental Magnetism. Technical Guide No. 6. Quaternary Research Association: London.