The global potential of log-driven trees for reconstructing forest ecosystems dynamics

1 Introduction

Over the past several centuries, lumber industries have been a crucial element in the economic development of many countries. In numerous regions, early timber felling lacked access to the few existing roads and railways for transporting logs (Barros and Uhl, 1995). Without this infrastructure, timber transport commonly relied on waterways (Agnoletti, 1995; Haba and Deusa, 2001; Törnlund and Östlund, 2002). This log drive served—for more than 200 years in some regions—as the primary means of moving timber from the forest to sawmills. Despite the importance of this historical disturbance, no study has reviewed where and when log driving was used.

Because of earlier intensive logging, many modern forest landscapes exist beyond their natural range of variability (Kuuluvainen et al., 2021). The current simplification and homogenization of forests make them highly vulnerable to climate change and its expected impacts on natural disturbance regimes (Montoro Girona et al., 2023a). Recently, forest management approaches have promoted harvesting techniques that emulate natural disturbance effects to restore and maintain forest structure and function (Stockdale et al., 2016). This ecosystem-based management requires knowledge of regional past disturbance regimes and preindustrial forest conditions (Gauthier et al., 2023; Montoro Girona et al., 2023a). However, despite current efforts to understand past forest states (Montoro Girona et al., 2018; Navarro et al., 2018; Aakala et al., 2023), there remains a significant knowledge gap in regards to past climate variations, disturbance regimes, and the state of preindustrial forests (Montoro Girona et al., 2023b). This missing information is critical for establishing more sustainable harvesting practices (Hof et al., 2021; Achim et al., 2022; Montoro Girona et al., 2023c).

During log driving, approximately 15% (eastern Canada) to 50% (central Siberia) of logs entering the waterways sank, resulting in often massive accumulations of wood on lake and river bottoms (Figure 1A; Marchand and Filion, 2014; Hellmann et al., 2016). Over the last decades, restoration efforts have aimed to return rivers and lakes to their natural state and function (Helfield et al., 2007; Pilotto et al., 2018). This restoration includes the removal of lost logs that, in some cases like in eastern Canada, have spent more than 200 years on the lake bottom (Tremblay, 1991; Houde, 2007). These sunken logs often remain in excellent condition (Figure 1B), enabling their recovery and commercial use (Lemay, 2016). These well-preserved subfossil logs are also of high interest for use in disturbance ecology as well as for reconstructing preindustrial forest ecosystems and the spatiotemporal patterns of forest exploitation.

Figure 1 (A) Subfossil logs on a lake bottom; the logs sank during log driving in the Mauricie region of Québec (Canada); (B) well-preserved logs extracted from Tee Lake (Témiscamingue, Canada); (C) stamp of the R.A. Hurdman Company on a recovered log (Ottawa, Ontario, ca. 1892); (D) stamp of the Grès Falls Company on a recovered log (Trois-Rivières, Québec, ca. 1908); (E) cross-section of a log recovered from a lake bottom, showing two fire scars. The green arrow indicates the first (oldest) fire, and the yellow arrow indicates the second (more recent) fire. Photo credits: (A) Nathalie Lasselin Aquanat, (B–D) Julie-Pascale Labrecque-Foy, and (E) Amélie Bergeron.

This review provides a global overview of log-driving history and highlights the significant value of submerged log-driven wood for disturbance ecology and the reconstruction of climate and preindustrial forest structure. We reviewed scientific (22) and historical (14) papers that mentioned the use of log driving and summarized where, when, and how log driving was used in multiple regions around the world (Table 1). We discuss how well-preserved submerged wood can help reconstruct the structure and dynamics of preindustrial forest ecosystems and provide relevant information for new management strategies to help protect and restore global forests.

Table 1 References stating the beginning and ending years of log driving in the different regions and countries around the world.

2 What is log driving?

Log driving is the transport of wood logs using river flow (Sedell et al., 1991). Cut logs are floated from the forest to main water bodies for direct export (Hardy, 2001) or, more frequently, to sawmills for processing (Sedell et al., 1991; Grabner et al., 2004). In most locations, log driving consisted traditionally of numerous steps, including marking, dumping, booming, raft building, transport, sorting, and storage. First, logs were marked on land at their cutting site by carving or stamping the wood with the company name or logo (Figures 1C, D) before being dumped into a nearby river (Schulman, 1974; Agnoletti, 1995). Dumping is the action of moving a log from the land to the water, either by rolling the log or by using slides, self-dumping barges, or even helicopters (White, 2001). Once in the water, and before their transport, the logs were kept in place by log booms—a barrier generally made of a chain of logs tied together by their ends (Schulman, 1974; White, 2001)—to immobilize the timber before being moved downstream. Transport of the logs was preferentially done during the high-water season (usually the spring freshet), when water levels and discharge are highest (Törnlund and Östlund, 2002). Logs could be floated individually or as bundles. When single logs were floated, they were guided by log drivers, i.e., workers standing on the logs or along the river bank, moving downstream with the logs and trying to prevent log jams (Pomerleau, 1997). Logs could also be tied together to build flat rafts (Grabner et al., 2021) to reduce the risk of logs sinking or drifting away (Hellmann et al., 2016). In certain rare cases, on larger rivers and when logs needed to be transported over a very long distance (sometimes up to 1500 km), rafts and wood bundles could be towed by boats (Adams, 1971; Barros and Uhl, 1995). Upon their arrival at the sawmill, the logs were sorted according to the company’s stamp as well as by grade, species, and size (Agnoletti, 1995; White, 2001). Logs were then stored in shallow water near the shoreline until they could be processed (White, 2001).

3 History of log driving around the world

Although log driving began earlier in some regions, log driving has been used on all continents and simultaneously across all regions during the 19^th and 20^th centuries (Figure 2).

Figure 2 Spatiotemporal pattern of log driving around the world.

3.1 Europe

Log driving first began in southern Europe, already being used by the end of the 10^th century on the Guadalquivir River in Andalusia (Spain). Timber companies floated mostly black pine (Pinus nigra) from the Cazorla and Segura forest for its use in the construction of palaces, religious buildings and ships (Haba and Deusa, 2001; Domínguez-Delmás et al., 2018). In Italy, log driving on the Piave River started in the 13^th century to bring wood from the Alps to Venice and lasted until around the end of the First World War (Agnoletti, 1995). In Austria, log driving began in the mid-16^th century and continued until the end of the Second World War (Grabner et al., 2004). The logs, mainly Norway spruce (Picea abies), were transported as rafts from the Limestone Alps to Vienna, and these rafts even served to transport passengers until 1860 (Grabner et al., 2021). In France, between the 16^th and the 19^th century, log driving was conducted on the Yonne river, from the Morvan mountainous massif to Paris (Jacod-Rousseau and Gob, 2020). Log driving was also used in France during the First World War, to supply wood to the British Army (Morris, 1920).

Log driving in Sweden developed at the beginning of the 19^th century (Björklund, 1984) and was followed by the extensive clearing of rivers and the construction of a network of floatways affecting more than 30,000 km of rivers and lakes (Törnlund and Östlund, 2002; Nilsson et al., 2005). Intensive exploitation of the forests started around 1850 (Nilsson et al., 2005), and by the end of the 19^th century, Sweden was a world leader in the exploitation of wood products (Björklund, 1984; Östlund, 1995). In Finland, the logging industry expanded quickly around 1860 (Björklund, 1984), and floatways available for log driving involved about 40,000 km of rivers (Nilsson et al., 2005). Over the second half of the 20^th century, timber trucks gradually replaced log driving, leading to the end of log driving by the end of the 20^th century in Scandinavia (Törnlund and Östlund, 2002; Hänninen, 2009).

3.2 Russia

The onset of log driving in Russian forests in the late 19^th century occurred mostly in the Baltic Sea area, although it expanded rapidly to the White Sea region (Björklund, 1984; Hellmann et al., 2016). The Pechora River later became the main waterway used for log driving (Hellmann et al., 2016). Log drive was also done along the Vistule river, from Russia to Poland (Daheur, 2019). The primary purpose for wood exploitation in Russia was exportation and, by the onset of the 20^th century, Russia was one of the largest wood exporters in Europe (Björklund, 1984). In 1965, about 125 million m³ of wood was estimated to be delivered by waterways in Russia, 80% transported as rafts (Avakyan, 1967). Log driving continues today in some areas of central Siberia and western Russia, although wood bundles and rafts are guided by boats (Hellmann et al., 2016).

3.3 North America

In the 19^th century, North American colonies were Britain’s most important wood suppliers (Aird, 2016). Commercial forestry and log driving in Québec (Canada) started in 1806 in the St. Lawrence River watershed and included the Québec City, Lac-Saint-Jean, Rimouski, Mauricie, and Outaouais regions (Pomerleau, 1997; Boucher et al., 2009; Gagnon et al., 2009; Chabot and Darveau, 2011; Marchand and Filion, 2014). In 1842, forest harvesting, accompanied by log driving, reached the Beauce–Appalachian region (Pomerleau, 1997) and then moved north to the Témiscamingue region by 1872 (Lienert, 1966). Timber harvesting also occurred in Ontario, from where the logs were floated on the Ottawa and St. Lawrence rivers before reaching Quebec City and being exported to Britain (Hardy, 2001). In most regions of eastern Canada, log driving ended around 1970 (Marchand and Filion, 2014).

In the United States, log driving began in multiple states between 1830 and 1870 (Sedell et al., 1991). In Maine, log driving occurred on many rivers between 1830 and 1976 (Moring et al., 1989). The first log drive in Pennsylvania took place in 1850 and was followed by intense logging (Cox, 1980), with large flat rafts of logs filling the western branch of the Susquehanna River in spring. In Kentucky and Tennessee, the Cumberland River was extensively used for log driving from 1870 to 1930 (Montell, 1988). Along this river, rafts containing up to 100,000 logs were floated down the river twice a year, once at the beginning of the winter and a second at the end of spring (Schulman, 1974). Log driving was also extensively used in the western United States. Rivers and streams in Nevada, Arizona, Montana, Utah, Colorado, Idaho, Oregon, Washington, and California all supported log drives at some point in the 19^th and 20^th centuries (Sedell et al., 1991).

Along the west coast of Alaska and British Columbia (Canada), log driving only began in the 20^th century (Pease, 1974; Sedell et al., 1991). Log driving continued in areas of British Columbia into the 21^st century (White, 2001).

3.4 South America, Asia, and Africa

Although less documented, log driving served as an important means of wood transport in South America, Asia, and Africa (De ’Ath, 1992; Barros and Uhl, 1995; Okon et al., 2021). In the Brazilian Amazon, commercial logging and log driving started around the beginning of the 17^th century. By the end of the 20^th century, log driving continued to be the main method of wood transport in remote areas (Barros and Uhl, 1995). In Thailand, between 1870 and 1937, teak logs were floated for 200 to 4000 km to reach Bangkok for export to Asia, Europe, and America (De ’Ath, 1992). Elephants were often used to move these logs from the cutting site to the river and to break log jams (De ’Ath, 1992). Finally, log driving continues to be practiced in Nigeria, as it remains the only option in regions with underdeveloped road networks (Okon et al., 2021).

4 The potential of sunken wood in forest ecology

4.1 Dendrochronology and forest ecology

Dendrochronology, i.e., the study of tree-ring records, has been widely used to reconstruct past climate (Pauly et al., 2021; Savard and Bégin, 2021) and disturbance regimes (Montoro Girona et al., 2016; Labrecque-Foy et al., 2020). Disturbances leave indicators such as scars, growth release, ring wedging, and growth suppression in trees’ annual radial growth rings (Swetnam et al., 2009; Morin et al., 2010; Debaly et al., 2022). Therefore, tree rings from living trees constitute valuable high-resolution environmental archives for reconstructing frequency, intensity, and spatiotemporal patterns of past disturbance regimes over the trees’ lifetimes (Pearson et al., 2014).

However, the temporal perspective of tree-ring analyses is limited by the life span of living trees (Stockdale et al., 2016). To increase the temporal perspective of tree-ring studies, researchers can also rely on wood collected from old buildings or dead trees (Krause, 1997; Payette et al., 2008; Boulanger et al., 2012). Dead wood found on land is often not well preserved, often preventing the development of long tree ring–based chronologies (Stockdale et al., 2016). However, accumulated wood preserved on lake bottoms can offer centennial- to millennial-scale records for boreal regions (Gennaretti et al., 2014). This preservation occurs in lakes having low lake-bottom oxygen levels and cold temperatures. These conditions limit decomposition (Pearson et al., 2014) and favor little to no modification of wood structure and anatomy (Beldean and Timar, 2021). Tree-ring chronologies from this well-preserved submerged wood, also called subfossil wood, can be dated when they overlap known and already-established chronologies (McCarroll and Loader, 2004). Therefore, subfossil wood enables researchers to study environmental conditions over a more extended period (Pearson et al., 2014). Given that 15% to 50% of log-driven logs sank during transport, the potential for such a paleoenvironmental resource is great for many boreal and mountainous regions.

4.2 Reconstruction of disturbance regimes

Fire is one of the main disturbances in the boreal forest, and the study of past fire regimes and its relation to climate and anthropic activities is necessary to improve forest management, particularly in the context of climate change (Aakala et al., 2023; Montoro Girona et al., 2023b). Trees at a fire’s edge or those growing within areas that experienced a moderate fire severity can develop fire scars (Erni et al., 2017). On a cross-section of a tree stem, fire scars can be identified (Figure 1E) and dated, even at a seasonal resolution (e.g., Swetnam et al., 2009), using the tree-ring record (Payette et al., 2008; Arseneault et al., 2021). Using fire scars from living and dead trees, Arseneault et al. (2021) reconstructed the fire dynamics over the last 200 years in northern Québec (Canada). However, using fires scars of subfossil wood extracted from six lakes in northern Québec, Gennaretti et al. (2014) were able to reconstruct a fire history covering about 1240 years.

Insect outbreaks can also be reconstructed with dendrochronology. For instance, during outbreaks of defoliating insects such as pine processionary moth (PPM; Thaumetopoea pityocampa) or spruce budworm (SBW; Choristoneura fumiferana), the defoliation of hosts trees leads to growth reductions that are noticeable in growth rings series (Morin et al., 2010; Sangüesa-Barreda et al., 2014). By dating radial growth reduction in living hosts species and comparing them to non-host species, researchers were able to reconstruct the past SBW outbreaks of the last two centuries in eastern Canada (Simard et al., 2011; Navarro et al., 2018). However, using subfossil wood, Pitre (2019) was able to reconstruct outbreaks episodes over the last 700 years, the longest chronology for SBW in North America.

4.3 Reconstruction of climate fluctuations

Climate reconstructions are essential for understanding how climate change affects forest ecosystems. Tree ring-based climate reconstructions rely on various properties, such as ring width, wood anatomy, and wood density (Bouriaud et al., 2005; Pritzkow et al., 2014; Büntgen et al., 2021). However, such records require detrending or standardization, which can sometimes modify or reduce the strength of the climatic signal (Labrecque-Foy et al., 2023). As an alternative, the isotopic analysis of the wood in the annual growth rings is receiving greater interest as a promising means of reconstructing past climate (Alvarez et al., 2018).

Oxygen and carbon stable isotope ratios (δ¹⁸O and δ¹³C, respectively) are influenced by photosynthesis rates and evapotranspiration, both influenced by environmental and climatic conditions (McCarroll and Loader, 2004). Combining oxygen and carbon isotopes of tree-ring records in wood beams from ancient buildings in Rennes (France), Masson-Delmotte et al. (2005) reconstructed summer temperatures and water stress for the last 400 years. However, using ¹⁸O isotopes of subfossil wood from one lake in northeastern Canada, Naulier et al. (2015) were able to reconstruct summer temperatures of the last 1000 years, and determined that the last three decades experienced the fastest warming over the last millennium.

4.4 Reconstruction of preindustrial forest composition and structure

Knowledge related to the composition and structure of preindustrial forest ecosystems is essential to set management goals and provide a reference state for future forest planning (Dupuis et al., 2011). The most direct means of assessing the natural state of preindustrial forests is to characterize old, never-harvested forests (Delisle-Boulianne et al., 2011). However, such forests are rare or non-existent in many regions because of earlier intense harvesting (Delisle-Boulianne et al., 2011). Therefore, researchers can rely on old land surveys to reconstruct past forest composition (Boucher et al., 2009; Terauds et al., 2011; Terrail et al., 2019). Although this technique has proven successful in specific contexts, the limited accessibility of such surveys, their inconsistent and low spatial coverage, and the relatively recent use of land surveys (often after the beginning of colonization) limit this approach (Boucher et al., 2014). Furthermore, old land surveys can be biased because they generally identify the dominant species, and rarely include the relative contribution of multiple subdominant species (Terauds et al., 2011).

Trees cut during the initial colonization phase represent preindustrial forests. Studying the log-driven subfossil wood of these preindustrial forests can help reconstruct the preindustrial forest composition and age. Boucher et al. (2009) combined old land surveys and subfossil wood to reconstruct past forest of the Rimouski region (Québec, Canada). Although selective logging could potentially bias a reconstitution, the authors found a good temporal representation of each study species. They reconstructed the state of the preindustrial forest (ca. 1820) and its gradual change in composition to the present.

The nature (size, species, age, etc.) of the wood found in the recovered logs can also inform about the practices of lumber industries over time (Boucher et al., 2009). They can help interpret how anthropic activities modified the composition and structure of forest ecosystem in the past (Jacod-Rousseau and Gob, 2020). For example, in the province of Québec (eastern Canada) selective logging occurred during the 19^th and 20^th century, which is reflected in the subfossil log-driven logs (Boucher et al., 2009). This is expected to be one of the main reasons why eastern white pines (Pinus strobus Linnaeus) and red pines (Pinus resinosa Ait) abundance significantly decreased in the forests during the 19^th century (Marchand, 2013; Danneyrolles et al., 2016). Sarting in the 20^th century, the transition from old conifer-dominated forests to young deciduous forests (Danneyrolles et al., 2019; Danneyrolles et al., 2021) is thought to be related to the end of selective logging and the extensive logging of numerous other conifer species such as spruce and fir (Barrette and Bélanger, 2007; Boucher et al., 2014).

4.5 Study of large wood dynamics in rivers

Forests and rivers are intimately connected, and the wood regime in rivers significantly influences river morphology (Wohl et al., 2019; Grosbois et al., 2023). Because log driving left persistent changes in river morphology (Jacod-Rousseau, 2022), efforts to restore rivers and their natural wood regimes gained interest in recent years (Wohl et al., 2019). Log-driven logs have great potential to inform about large wood debris dynamics in rivers. Indeed, the location of the logs can help identify sites more suitable for wood accumulation and their characteristics, and target the size (diameter and length) and species of logs that preferentially accumulate in rivers (Riuz-Villanueva et al., 2016). The level of decay of the logs and the dating of their last growth ring by dendrochronology can help inform about the residence time in the river (Riuz-Villanueva et al., 2016). Furthermore, the volume and biomass of log-driven logs that remain in the waterways of the boreal forests have not yet been determined. These information are crucial to restore the natural wood dynamics in rivers worldwide and to preserve aquatic and riparian habitats (Wohl et al., 2019).

5 Conclusion

Log driving played a major role in early forestry practices in many countries; however, the historical record of log driving remains poorly known. We note that log driving has been used on most continents at some point in history, and simultaneously in numerous countries during the 19^th and 20^th centuries. However, there exist a large bias favoring North America and Eurasia in the available literature. Also, historical documents were difficult to access, and their exploration can be limited by a language barrier and state of conservation, leaving a knowledge gap that can limit our understanding of this practice and its broader implications around the world.

Although intensive logging has had major negative impacts on forest landscapes, the logs that sank to lake bottoms during log driving also represent an invaluable resource for reconstructing past disturbances, climate, and the composition of preindustrial forest ecosystems. In the context of an ever-increasing demand for wood and global warming, this longer-term knowledge derived from log-driven wood is essential for improving forest management approaches in the face of climate change.

Author contributions

Conceptualization: J-PL-F and MMG. Writing—original draft: J-PL-F. Writing—review and editing: J-PL-F and MMG. Both authors contributed to the article and approved the submitted version.

Funding

This research was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC) discovery grant to reconstruct the regime of natural disturbances in forest ecosystems (RGPIN-2022-05423) obtained by MG. J-PL-F was supported by the excellence scholarship programs from Vanier Canada Graduate Scholarships (CGV-180844) and the Fonds de Recherches du Québec – Nature et Technologies.

Acknowledgments

The authors are thankful to the Mauricie National Park team compilating information about the log drive history, to Amélie Bergeron and Nathalie Lasselin for some of the pictures and to Murray Hay for linguistic revision.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References