Abstract

The growing field of historical ecology describes population abundances, biodiversity, spatial structure, and ecological functioning in the past, which are important to understanding ecological dynamics and recovery potential. However, because historical ecological analyses often incorporate nontraditional data sources (i.e., archival materials and oral histories) and may reveal unexpected changes to species and populations, the results are subject to critiques of objectivity and quality control, which may hamper their broad application. Here, we argue that surprising results revealed by historical sources are essential for ecology and conservation, providing new hypotheses that can be tested with additional data and new understandings of ecological dynamics that have immediate conservation implications. We outline four best practices in conducting conservation-relevant historical ecology research using nontraditional sources, and we argue that if these practices are considered in designing historical ecological analyses, the results can provide reliable insight into past change, current ecosystem structure, and future recovery targets.

Historical ecology developed as an organized research approach in the middle of the twentieth century (Szabó 2014), arising from a convergence of ecological anthropologists who found that systems thinking provided insight into the long-term coupling of human societies and their impacts on the environment (Balée 2006) and ecologists who were starting to formally analyze the effects of prior human impacts on a diverse array of ecosystems (e.g., Swetnam et al. 1999). Historical ecology research programs have been unified by a core belief that understanding present biotic conditions requires viewing them through the lens of past interactions with human societies (Balée 2006). Since the 1990s, global interest in historical ecology has grown (Szabó 2014), with expansion into marine systems (Jackson et al. 2001, Kittinger et al. 2015). Modern historical ecology intersects strongly with environmental history, ecological anthropology, historical geography, and paleoecology, with researchers from all disciplines contributing to the understanding of long-term change (Balée 2006, Szabo 2014). Whereas anthropological and geographical approaches to historical ecology focus on patterns in processes in landscapes (Balée 2006), paleoecological and ecological approaches also include ecosystem-level changes, such as the indirect effects of species’ removals on ecosystem function (Jackson et al. 2001). Results from both approaches have been essential to biodiversity conservation, with strong links to restoration ecology (Szabó and Hédl 2011), as well as understandings of biological invasions, frequency of disturbance, ecological phase shifts, and the natural variability of ecosystem state and population abundances (Jackson et al. 2001, Willis and Birks 2006).

Historical ecologists have relied on an integration of diverse sources, including physical evidence, such as archaeological remains, pollen records, and genetic analyses, and sources more typical of the humanities and social science, such as archival materials and oral histories. Recent research has documented changes to population abundances, metapopulation dynamics, and ecosystem structure and function using a variety of sources, including maps (Walter and Merritts 2008), written anecdotes (McClenachan and Cooper 2008), oral histories (Ames 2004), photographs (McClenachan 2009), newspaper articles (Thurstan et al. 2014), and even restaurant menus (Van Houtan et al. 2013). These nontraditional sources often fill crucial data gaps (McClenachan et al. 2012), providing insight into change on the scale of decades to centuries, often during time periods in which human impacts accelerated but before ecological data collection began (figure 1).

Figure 1.

A typical example of types of sources available through time. Historical data sources fill an important temporal gap, corresponding with periods during which human impacts (e.g., habitat alteration and species exploitation) accelerated but before ecological data collection began.

Figure 1.

A typical example of types of sources available through time. Historical data sources fill an important temporal gap, corresponding with periods during which human impacts (e.g., habitat alteration and species exploitation) accelerated but before ecological data collection began.

Although the use of archival materials and oral histories has a long tradition in historical ecology (Szabo 2014), these sources are often considered to be nontraditional, particularly in marine systems, where historical ecology research is more recent. As a result, the use of these sources has invited criticism, particularly when the results contradict existing scientific knowledge. Here, we argue that the sometimes surprising results revealed by historical sources are essential for both ecology and conservation, providing new hypotheses that can be tested with additional data and new understandings of ecological dynamics that have immediate conservation implications. We describe four ways in which ecologists using historical data can use best practices to ensure the reliability of the results, taking lessons from other disciplines, including history, archaeology, and paleoecology; these include (1) identify a full range of sources relevant to the research question, (2) address biases in the historical record, (3) design analyses to deal with data gaps, and (4) use multiple lines of evidence while understanding the limitations of each. Our goals are to discuss the necessary elements of high-quality research for ecologists wishing to use nontraditional data to analyze past change, to provide examples of pitfalls and successes, and to discuss the novel insights that have emerged from their use. We draw on a variety of examples from terrestrial, freshwater, and marine systems. However, because the majority of reviews of historical ecology literature to date have focused on terrestrial systems (Sheail 1980, Swetnam et al. 1999, Edmunds 2005, Balée 2006, Szabó 2014, Szabó and Hédl 2011), we rely more heavily on the recent and growing literature in marine historical ecology.

Surprising results revealed by historical sources

Historical sources have revealed surprising results that can challenge established scientific ideas about species or ecosystems, providing opportunities for new hypotheses and ultimately new understandings of ecological dynamics. Such surprising results can have immediate conservation implications across a range of species and ecosystems (figure 2). In particular, our understanding of species’ past and present ranges is frequently incomplete and can benefit from the use of historical data. For example, the highly endangered vaquita (Phocoena sinus) was assumed to naturally inhabit a small area in Mexico's northern Gulf of California, but drawings made by eighteenth-century explorers identified individuals in more than 1000 kilometers outside of this assumed range (Sáenz Arroyo et al. 2006). Likewise, early reports from North American fur trappers indicated that beaver (Castor canadensis) historically existed in areas where scholars had not recorded it (Lanman et al. 2012). In the Atlantic Ocean, an analysis of historical whaling logbooks suggested that the feeding range of humpback whales (Megaptera novaeangliae) extended to the Mid-Atlantic Ridge, far outside of the known range at the time (Reeves et al. 2004). Oral histories can reveal similar hidden changes, such as those for Atlantic cod (Gadus morhua), which were demonstrated to have lost fine-scale population structure in near-shore areas of the Gulf of Maine over the past century because of intensive fishing (Ames 2004).

Figure 2.

Species for which surprising historical data altered the conservation outcome: A. Humpback whale (Megaptera novaeangliae), B. North American beaver (Castor canadensis), C. Atlantic cod (Gadus morhua), D. balloon wine (Cardiospermum corindum). Photos from Wikicommons.

Figure 2.

Species for which surprising historical data altered the conservation outcome: A. Humpback whale (Megaptera novaeangliae), B. North American beaver (Castor canadensis), C. Atlantic cod (Gadus morhua), D. balloon wine (Cardiospermum corindum). Photos from Wikicommons.

In each of these cases, insights gained from historical sources challenged modern assumptions about species’ natural distribution, providing opportunities for new hypotheses to be tested with additional data. The use of radiocarbon dating of wood excavated from recently exposed ancient beaver dams confirmed that the range of beavers extended into California's Sierra Nevada before 1850, which proved important information to the management of rapidly expanding modern beaver populations (Lanman et al. 2012). Because of its ecosystem engineering, the beaver's historical presence has multiple ecological implications that influence our understanding of natural ecosystem structure and function, including the structural stability of streams, nutrient fluxes, and fish diversity. The results of the whaling logbook analysis prompted satellite tracking of extant individuals, which confirmed that humpback whales use a larger area than had previously been known. Together, these two sets of data provided justification for the expansion of protection, because they indicated that whales actively use an area broader than the existing sanctuaries (Kennedy et al. 2013). Genetic analyses of Atlantic cod in the Gulf of Maine supported oral history accounts, demonstrating that populations are structured on a finer scale than previously thought. As a result, discussions have begun about how to tailor management plans to account for this greater biocomplexity and locally adapted populations (Kovach et al. 2010).

Historical data can also fundamentally alter our understanding of ecosystem dynamics, including the physical structure of ecosystems, ecosystem function, species interactions, and phase shifts, each with important implications for restoration goals and conservation strategies. Walter and Merrits (2008) used historical maps and other records to demonstrate that before European settlement, streams in the eastern United States formed small channels within extensive vegetated wetlands, a structure that was facilitated by abundant beaver populations. These findings contradicted the assumptions about natural meandering stream structure that had previously guided a multibillion-dollar stream restoration industry (Bernhardt et al. 2005). The decline of oyster reefs in estuaries is well known, but the structure and function of historical reefs is difficult to measure. Zu Ermgassen and colleagues (2012) used nineteenth-century surveys to demonstrate that whereas the area of oyster reefs in the United States declined by 64% over the past century, their density declined by nearly 90%. The difference between these two numbers suggested that historical and modern oysters reefs are structurally different, with historical reefs subsequently more efficient at water filtration, an essential ecosystem service. These findings indicated that restoration targets should consider ecosystem function as a separate and an essential metric of success in restoring oyster reefs (Kittinger et al. 2015). Understanding invasion ecology—and the relationships among species before human impact—has been an important contribution of historical ecology (Willis and Birks 2006). In south Florida, historical data showed that a plant species previously considered an invasive weed, balloon wine (Cardiospermum corindum), was actually a native species, which provided crucial habitat for an endangered butterfly (Carroll and Loye 2006), thereby altering the conservation outcome for this species. In the Caribbean Sea, an observed phase shift from coral- to algal-dominated benthic structure was proximally linked to a disease that eliminated the herbivorous urchin (Diadema antillarum), but historical data on the intensity of past fishing and fish abundance suggested the ultimate driver was the loss of functional redundancy provided by abundant herbivorous fish (Jackson et al. 2001). These findings, along with more recent observations, provided justification for the increased protection of grazing fishes such as parrotfishes (Jackson et al. 2014). These conservation-relevant results all stemmed from historical data that challenged previously accepted scientific knowledge. As such, they may have been dismissed but instead revealed important and fundamental changes that would have been overlooked across a range of species and ecosystems.

Best practices in historical ecology

Although surprising results revealed by historical sources can provide important new ecological insights with strong implications for conservation, there are several potential limitations involved with using historical data for ecological analyses. To be reliable, we argue, there are several best practices in historical ecology. These include identifying a full range of sources relevant to the research question, addressing biases in the historical record, designing analyses to deal with data gaps, and using multiple lines of evidence while understanding the limitations of each. In our discussion of these topics, we draw on literature about the use of documentary sources and oral histories in historical ecology (Sheail 1980, Drew 2005, Edmunds 2005, Szabó and Hédl 2011), enhanced with recent examples of research that has employed these best practices.

Identify a full suite of sources relevant to the research question

Historical ecological research begins with region, ecosystem or species of interest, and a research question that addresses past human actions and long-term ecological change (Szabó and Hédl 2011). Establishing clear goals at the start of the project is important not only for the integrity of the research but also for setting the boundaries on the kinds of data that are relevant. In particular, understanding the timing of key drivers of change will help to understand the timeframe of historical data required, which allows the researcher to narrow in on key types of sources (figure 1). Within the delineated time frame, identifying a full suite of sources relevant to the research question is essential, because omitting key types of sources can substantially alter the research conclusions. Similar to archaeology and paleoecology, historical ecology must draw conclusions from existing evidence rather than generate new experimental data to test hypotheses; therefore, before analyses begin, this evidence must be as complete as possible. Here, we discuss issues related to identifying relevant archival sources and Traditional Ecological Knowledge (TEK).

Finding archival sources involves understanding the motivations of people in the past, the types of records they kept, and the multiple locations where these records may be maintained, including public archives, libraries, and museums or private collections (Sheail 1980, Egan 2005, Szabo and Hedl 2011). Because of data availability, the majority of historical ecology research has focused on change over the past 200 years, but human impact on ecosystems often extends much earlier. Therefore, the researcher must be aware of the limitations of available data in fully capturing long-term change (Szabó and Hédl 2011). In many regions of the world, the earliest written records include narrative descriptions, nautical charts, and land survey records ­(figure 1). Particularly where they correspond with exploration and colonization, these sources often describe species with taxonomical and geographic precision, including foundation species such as trees and reef-building oysters (Zu Ermgassen et al. 2012) and large animals, many of which are now conservation dependent (McClenachan et al. 2006). As trade networks developed and intensified, records documenting exploitation and trade increased in frequency, such as customs records and logbooks. These sources often provide the earliest quantitative data that can be used to infer species’ abundances (e.g., Rosenberg et al. 2005). From the twentieth century, a far greater variety of sources becomes available, including photographs and newspaper articles, from which levels of exploitation and changes in abundance may be quantified (McClenachan 2009, Thurstan et al. 2014). Although many of these sources may be dismissed as anecdotal, written descriptions of the locations of natural phenomena or descriptions of particular species as “rare” or “abundant” may provide key information about past change. Indeed, compilations of large numbers of historical anecdotes from diverse sources have formed the basis of a range of analyses, including spatial changes in population structure (McClenachan and Cooper 2008), changes in the relative abundance of functional groups in coral reef ecosystems over centuries (Pandolfi et al. 2003), and assessments of historical ecosystem recovery (Kittinger et al. 2011).

In some cases, information about early population dynamics may exist only in the living memory of individuals, such that oral histories may be essential to capture long-term change. This collected, multigenerational knowledge held by individuals living within specific localities, or TEK, can offer insight into ecological interactions, species distributions, and culturally salient species that may not be either readily apparent or even still extant (Drew 2005). For example, Burney and Ramilisonina (1998) described how Malagasy folk stories recount ecological and behavioral information about extinct megafauna in Madagascar. Likewise, interviews with generations of resource users demonstrated changes in relative abundance, size structure, and population structure for the Gulf grouper (Mycteroperca jordani) in the Gulf of California for decades where no other data existed (Sáenz Arroyo et al. 2005). TEK is important for understanding which species were (and are) important to the lives of people living in an area and, by extension, which species are likely to be involved with human species interactions. As with many other collaborative research efforts, research in TEK requires working across disciplinary boundaries, and the resulting data may have different meanings to researchers with different epistemological backgrounds (Szabó and Hédl 2011).

Understand biases in the historical record

Interpreting historical documents requires understanding the context in which the documents were created and preserved and recognizing what may have gone unrecorded (Sheail 1980, Edmunds 2005). In the field of history, source criticism—the analysis of what was recorded, why, by whom, and for what reasons—forms an integral and foundational stage in historical analysis (Szabó and Hédl 2011). Similarly, biases in the historical record influence the types of information available and determine the types of ecological conclusions that can be reached. These include observation bias, sampling effort, preservation bias, and recording bias, all of which depend on the cultural context in which the observations were made (table 1).

Table 1.

Types of biases in historical records and implications for historical ecological analyses.

Type of bias Description Implication for analyses 
Observation bias Observations of species are biased toward those with economic or cultural value; Common species are less often recorded than rare species Abundance cannot be inferred from the absolute number of records or observations; Species records cannot be used to quantify changes in biodiversity 
Sampling effort Numbers of observations are rarely consistent across time and space Sampling effort must be estimated to understand relative abundance; Gaps in data may reflect a lack of observation 
Preservation bias Original documents may be lost, fragmented or degraded Early descriptions should not be confused with direct observation; A lack of written data does not imply historical absence 
Recording bias Incentives or disincentives to record information exist Historical records (and the lack of records) must be interpreted in a cultural context 
Type of bias Description Implication for analyses 
Observation bias Observations of species are biased toward those with economic or cultural value; Common species are less often recorded than rare species Abundance cannot be inferred from the absolute number of records or observations; Species records cannot be used to quantify changes in biodiversity 
Sampling effort Numbers of observations are rarely consistent across time and space Sampling effort must be estimated to understand relative abundance; Gaps in data may reflect a lack of observation 
Preservation bias Original documents may be lost, fragmented or degraded Early descriptions should not be confused with direct observation; A lack of written data does not imply historical absence 
Recording bias Incentives or disincentives to record information exist Historical records (and the lack of records) must be interpreted in a cultural context 

Throughout history, people have written about what interests them and what they believe will interest their intended audience. Therefore, evaluating historical context first involves understanding the observation bias of the person who created the historical record (Sheail 1980, Edmunds 2005, Szabó and Hédl 2011). For example, although cannibalism was probably a very small portion of pre-Colonial Fijian diet (Jones et al. 2012), the Fiji Islands maintained an undeserved reputation for cannibalism in western literature for decades (Sahilins 2003). These narratives of the illicit were useful in prompting sales of travel literature. More importantly, by providing a context through which to justify both missionary activities and later colonialism in the region, they served as an avenue for dispossession (Banivanua-Mar 2010). Early European writing about species and ecosystems were similarly biased by their intended audiences, which included those sponsoring the voyages, fellow sailors, and the contemporary European public. As a result, species that were commercially valuable, edible, or unique were more likely to be documented (e.g., Reeves et al. 2004, Rosenberg et al. 2005). Paradoxically, common species were often least well documented, because they were simply not interesting enough to draw notice (Sheail 1980). This observation bias means that the absolute number of records on a particular species cannot be used directly as a proxy of abundance or rarity. Similarly, the bias in the historical record toward rare, exceptional, or commercially important species hampers the use of historical accounts of species to quantify changes in biodiversity. However, these records provide essential information on species that may be rare or extinct today and form the basis of analyses of historical exploitation, long-term population change, range shifts, and the ecological implications of these long-term changes, all of which would otherwise be unknown.

Understanding historical context also involves interpreting the sampling effort, both on an individual level (where did the observer travel?) and cumulatively (how many potential observers were there?). If sampling effort is not evaluated, there is a risk of incorrectly interpreting a lack of historical documentation as rarity of species. Therefore, uncalibrated mentions of particular species can provide an indication of presence but cannot be used to directly track abundance. However, if sampling effort can be estimated, it is possible to quantify relative abundance over time or across space. For example, Luiz and Edwards (2011) used 150 years of ships’ logs to estimate the timing of extinction of shark populations in this remote island system. This work involved a careful review of the historical sampling effort over this time period, including a review of early explorers visiting the area, fisheries in the midtwentieth century, and scientific sampling in recent decades.

A third element of historical context is preservation bias. In paleoecology and archaeology, some types of materials preserve better, longer, and more intact than others, which is the reason that our understanding of past mass extinction events is highly reliant on certain taxa such as marine invertebrates (Jablonski and Chaloner 1994). Similar taphonomic processes exist for historical documents, because a large proportion of historical documents have been destroyed (Sheail 1980). For example, Columbus’ accounts of his travels provide one of the first written descriptions of the New World and contain useful information about the state of the natural environment in the fifteenth century. However, Columbus's original logs have not survived, so what we know of these journeys was either copied from his originals, written by his officers, or extracted later by others (Cohen 1969). Therefore, although they provide useful baseline information about natural environments, such early descriptions should not be confused with direct observation, and researchers must be aware that they will never find all the source documents required to fully understand past events (Sheail 1980). As such, a lack of observation does not equal a historical absence, but an understanding of historical context can help understand the biases in what materials have been preserved or lost.

Finally, understanding recording bias and the cultural context in which historical documents were created is essential (Sheail 1980, Szabó and Hédl 2011). For example, Lanman and colleagues (2012) explained that the lack of written data about beaver in California was not due to their historical absence from the region but rather to the reticence of trappers in revealing exactly where profitable trapping grounds existed and to the disincentive to report pelts to government record keepers. Taken out of context, mammalogists interpreted this lack of data as a lack of abundance and concluded that the species was not found in high elevations. Similarly, Baisre (2013) concluded that the now-extinct Caribbean monk seal was rare at the time of European exploration because of an observation that they were consumed by elites among Native American communities in Florida. Although exclusive consumption could point to rarity, cultural prohibitions are often independent of species biology. For example, elite households in pre-contact Hawaii consumed significantly greater quantities of fish and chicken, but neither of these fauna was rare (Kirch and Jones O'Day 2003). These two examples illustrate how, without outside information to contextualize observations, data ­cannot be used to infer rarity or abundance.

Design analyses to deal with data gaps

The number of historical records may or may not coincide with abundance, commonness, or rarity. However, analyses can be designed to address data gaps. In this case, interpretation of historical sources can borrow from archaeological research, which commonly involves work with incomplete sources, and has developed techniques to control for source biases and information gaps. For example, if sampling effort is kept constant, comparisons across taxa can be used to infer relative abundance, such as from longer-lived to shorter-lived species in response to intensive exploitation (McClanahan and Omukoto 2011). Similar comparisons have also been effective in analyses of historical data provided that sampling effort is kept constant or calibrated, such as assessments of the relative abundance of species of trophy fish in historical photos (McClenachan 2009) and reef fish, turtles, and other local taxa on restaurant menus (Van Houtan et al. 2013).

Similarly, time-series data are typically lacking over long time scales, but if temporal data gaps exist, “then” and “now” comparisons can be made to quantify relative population change. Lotze and Milewski (2004) determined the relative abundance of seven species of marine mammals and fish in the Quoddy Region of the Bay of Fundy by making comparisons of catch or observed abundance at two points in time, whereas Zu Ermgassen and colleagues (2012) compared records of abundance between the late nineteenth and twentieth centuries to quantify the decline in oyster populations in the United States. Alexander and colleagues (2009) used fisheries’ logbook data to show that Atlantic cod landings in the Gulf of Maine were twenty times greater in 1861 than they are today. When only two data points are compared, it must be kept in mind that a single historical baseline does not encompass a full range of natural variability. However, these analyses are an essential step toward understanding the range of variability in ecosystems.

Even if it is not possible to quantify change in relative abundance, general trends can be identified from historical records. Broad patterns from spotty data may be better than no data, which implies that no change occurred. If those limits are kept at the forefront of the researchers’ minds, they may provide important insights to later periods that do have more detailed data. For example, in addition to quantifying changes to cod populations, Alexander and ­colleagues (2009) highlighted the potential for whole ecosystem change using fishermen's written descriptions of cod feeding behavior and logbook records documenting bait fish used in the fishery. Together these sources suggested a loss of forage fish from these ecosystems since the 1860s. In Hawaii and the Florida Keys, reconstructions of historical catch allowed for an understanding of historical fisheries and management structures over century-long time scales over which no other data existed (McClenachan and Kittinger 2012). In both of these cases, the broad patterns that emerged contribute to an overall understanding of change and the ways in which human activities have altered community structure, which otherwise might have been overlooked.

Use multiple lines of evidence and understand the limitations of each

Finally, because data availability is often limited, the use of multiple lines of evidence to design analyses and ground truth results is essential (Edmunds 2005). Information sources that work together to provide supporting lines of evidence can include either historical data and experimental findings or various lines of historical data. For example, a team of historians and fisheries biologists set out to quantify past population size for Atlantic cod, drawing on their knowledge of the nineteenth-century fisheries business environment to locate New England ships’ logs from the 1850s that contained daily cod catch records. Once they compiled these records, they matched them with shipping records kept in the US National Archives, which provided an estimate of vessel tonnage and crew size. Together, these two sets of documents provided robust catch-per-unit-effort data and allowed for the estimation of cod biomass removed from the Scotian Shelf a century before traditional fisheries data exist (Rosenberg et al. 2005). In the Caribbean, McClenachan and Cooper (2008) estimated historical population densities of Caribbean monk seals on the basis of historical trade data and then used food-web models to determine whether those estimates were ecologically possible. The results indicated that historical reefs could have sustained estimated seal populations if they contained approximately the same prey biomass as modern reference reefs in the Pacific, suggesting that the population estimates were reasonable.

As with any research, it is important to understand the limitations of historical ecology analyses. Analyses of range changes over time are typically based on simple observations of presence that require relatively little interpretation. Similarly, assessments of relative abundance of single species based on multiple anecdotes describing that species’ rarity or abundance over time have been shown to be robust (Al-Abdulrazzak et al. 2012). However, estimates of absolute abundance typically require additional evidence to parameterize models and data sets to confirm results. For example, Rosenberg and colleagues’ (2005) estimate of historical cod abundances based on logbook records agrees with previous estimates based on carrying capacity (Myers et al. 2011), suggesting that the model was well parameterized. However, estimates of humpback whale (Megaptera novaeanglia) populations based on logbook records are two to three times lower than estimates based on the genetic diversity of present-day populations (Ruegg et al. 2013), indicating that more research is required to account for the difference. Such analyses of contemporary DNA have proven useful in estimating past population sizes (e.g., Ruegg et al. 2013) and in estimating historical abundances of severely depleted or extinct species; analyses using ancient DNA could prove similarly useful. Likewise, ecosystem models that investigate how ecosystems could have functioned before the loss of top predators are strongest when corroborated with modern observations from reference ecosystems.

Conclusions

Historical ecology is a developing discipline that has made important ecological contributions, and the use of novel archival and oral history sources has filled temporal gaps between paleoecology and current ecological research. The growing use of nontraditional sources has helped to reduce data gaps and provided conservation relevant findings. Although the results are often surprising, dismissing them because of modern perceptions of “natural” ecological states can lead to the loss of information that could inform future research. Instead, such surprising results provide essential information for the advancement of ecological knowledge and the direction of conservation actions, in particular in generating new hypotheses that can be tested with additional data. However, these sources should not be used uncritically, and the use of best practices is needed to produce robust replicable analyses. By highlighting the benefits of being inclusive in the kinds of data incorporated and the importance of placing those data in a historical context, this approach expands the kinds of data that can help inform conservation. When ecologists incorporate data from nontraditional sources, we are able to explore past ecosystems, quantify historical baselines, and ultimately evaluate the magnitude and direction of past ecosystem change. It is with this knowledge that we will be able to evaluate how far we have come and where we are to go.

We are grateful to Ruth Thurstan and the three anonymous reviewers for their constructive feedback on earlier versions of this manuscript. LM was supported by an Alfred P. Sloan Research Fellowship (grant no. FG-BR2013-071).

References cited

Al-Abdulrazzak
D
Naidoo
R
Palomares
MLD
Pauly
D
Gaining perspective on what we've lost: The reliability of encoded anecdotes in historical ecology
PLOS ONE
 
2012
7
art. e43386
Alexander
KE
et al
Gulf of Maine cod in 1861: Historical analysis of fishery logbooks, with ecosystem implications
Fish and Fisheries
 
2009
10
428
449
Ames
EP.
Atlantic cod stock structure in the Gulf of Maine
Fisheries
 
2004
29
10
28
Baisre
J.
Shifting baselines and the extinction of the Caribbean monk seal
Conservation Biology
 
2013
27
927
935
Balée
W.
The research program of historical ecology
Annual Review of Anthropology
 
2006
35
75
98
Banivanua-Mar
T.
Cannibalism and colonialism: Charting colonies and frontiers in nineteenth-century Fiji
Comparative Studies in Society and History
 
2010
2
255
281
Bernhardt
ES
et al
Synthesizing US river restoration efforts
Science
 
2005
308
636
637
Burney
DA
Ramilisonina
The kilopilopitsofy, kidoky, and bokyboky: Accounts of strange animals from Belo-sur-mer, Madagascar, and the megafaunal “extinction window.”
American Anthropologist
 
1998
100
957
966
Carroll
SP
Loye
J
Invasion, colonization, and disturbance: Historical ecology of the endangered Miami blue butterfly
Journal of Insect Conservation
 
2006
10
13
27
Columbus
C.
Cohen JM, trans. The Four Voyages: Being His Own Log-Book, Letters, and Dispatches with Connecting Narratives
 
1969
Penguin Books
Drew
JA
Use of traditional ecological knowledge in marine conservation
Conservation Biology
 
2005
19
1286
1293
Edmunds
M.
Egan
D
Howell
EA
The pleasures and pitfalls of written records
The Historical Ecology Handbook: A Restorationist's Guide to Reference Ecosystems
 
2005
Island Press
73
100
Jablonski
D
Chaloner
WG
Extinctions in the fossil record
Philosophical Transactions of the Royal Society B
 
1994
344
11
17
Jackson
JBCJ
et al
Historical overfishing and the recent collapse of coastal ecosystems
Science
 
2001
293
629
637
Jackson
JBCJ
Donovan
M
Cramer
K
Lam
V
Status and Trends of Caribbean Coral Reefs: 1970–2012
 
2014
Global Coral Reef Monitoring Network
Jones
S
Walsh-Haney
H
Quinn
R
Kana Tamata or feasts of men: An interdisciplinary approach for identifying cannibalism in prehistoric Fiji
International Journal of Osteoarchaeology
 
2012
25
127
145
doi:10.1002/oa.2269
Kennedy
AS
Zerbini
AN
Vasquez
OV
Gandilhon
N
Clapham
PJ
Adam
O
Local and migratory movements of humpback whales (Megaptera novaeangliae) satellite-tracked in the North Atlantic Ocean
Canadian Journal of Zoology
 
2013
92
9
18
Kirch
P
Jones O'Day
SJ
New archaeological insights into food and status: A case study from pre-contact Hawaii
World Archaeology
 
2003
34
484
497
Kittinger
JN
Pandolfi
JM
Blodgett
JH
Hunt
TL
Jiang
H
Maly
K
McClenachan
L
Schultz
JK
Wilcox
BA
Historical reconstruction reveals recovery in Hawaiian coral reefs
PLOS ONE
 
2011
6
art. e25460
Kittinger
JN
McClenachan
L
Gedan
KB
Blight
LK
Marine Historical Ecology in Conservation: Applying the Past to Manage for the Future
 
2015
University of California Press
Kovach
AI
Breton
TS
Berlinsky
DL
Maceda
L
Wirgin
I
Fine-scale spatial and temporal genetic structure of Atlantic cod off the Atlantic coast of the USA
Marine Ecology Progress Series
 
2010
410
177
195
Lanman
RB
Perryman
H
Dolman
B
James
CD
The historical range of beaver in the Sierra Nevada: A review of the evidence
California Fish and Game
 
2012
98
65
80
Lotze
HK
Milewski
I
Two centuries of multiple human impacts and successive changes in a North American food web
Ecological Applications
 
2004
14
1428
1447
Luiz
OJ
Edwards
AJ
Extinction of a shark population in the Archipelago of Saint Paul's Rocks (equatorial Atlantic) inferred from the historical record
Biological Conservation
 
2011
144
2873
2881
McClanahan
TR
Omukoto
JO
Comparison of modern and historical fish catches (AD 750–1400) to inform goals for marine protected areas and sustainable fisheries
Conservation Biology
 
2011
25
945
55
McClenachan
L.
Documenting loss of large trophy fish from the Florida Keys with historical photographs
Conservation Biology
 
2009
23
636
643
McClenachan
L
Cooper
AB
Extinction rate, historical population structure, and ecological role of the Caribbean monk seal
Proceedings of the Royal Society B
 
2008
275
1351
1358
McClenachan
L
Jackson
JBCJ
Newman
MJH
Conservation implications of historic sea turtle nesting beach loss
Frontiers in Ecology and the Environment
 
2006
4
290
296
McClenachan
L
Ferretti
F
Baum
JK
From archives to conservation: Why historical data are needed to set baselines for marine animals and ecosystems
Conservation Letters
 
2012
5
349
359
McClenachan
L
Kittinger
JN
Multi-century trends and the sustainability of coral reef fisheries in Hawai'i and Florida
Fish and Fisheries
 
2013
14
239
255
Myers
RA
MacKenzie
BR
Bowen
KG
Barrowman
NJ
What is the carrying capacity for fish in the ocean? A meta-analysis of population dynamics of North Atlantic cod
Canadian Journal of Fisheries and Aquatic Sciences
 
2001
58
1464
1476
Pandolfi
JM
et al
Global trajectories of the long-term decline of coral reef ecosystems
Science
 
2003
301
955
958
Reeves
RR
Smith
TD
Josephson
EA
Clapham
RJ
Woolmer
G
Historical observations of humpback and blue whales in the North Atlantic Ocean: Clues to migratory routes and possible additional feeding grounds
Marine Mammal Science
 
2004
20
774
786
Rosenberg
AA
Bolster
WJ
Alexander
KE
Leavenworth
WB
Cooper
AB
McKenzie
MG
The history of ocean resources: Modeling cod biomass using historical records
Frontiers in Ecology and the Environment
 
2005
3
78
84
Ruegg
K
Rosenbaum
HC
Anderson
EC
Engel
M
Rothschild
A
Baker
CS
Palumbi
SR
Long-term population size of the North Atlantic humpback whale within the context of worldwide population structure
Conservation Genetics
 
2013
14
103
114
Sáenz-Arroyo
A
Roberts
CM
Torre
J
Cariño-Olvera
M
Enríquez-Andrade
RR
Rapidly shifting environmental baselines among fishers of the Gulf of California
Proceedings of the Royal Society B
 
2005
272
1957
1962
Sáenz Arroyo
A
Roberts
CM
Torre
J
Carino-Olvera
M
Hawkins
JP
The value of evidence about past abundance: Marine fauna of the Gulf of California through the eyes of sixteenth- to nineteenth century travellers
Fish and Fisheries
 
2006
7
128
146
Sahlins
M.
Artificially maintained controversies: Global warming and Fijian cannibalism
Anthropology Today
 
2003
19
3
5
Sheail
J.
Historical Ecology: The Documentary Evidence
 
1980
Institute of Terrestrial Ecology
Swetnam
TW
Allen
CD
Betancourt
JL
Applied historical ecology: Using the past to manage for the future
Ecological Applications
 
1999
9
1189
1206
Szabó
P.
Historical ecology: Past, present, and future
Biological Reviews
 
2014
doi:10.1111/brv.12141
Szabó
P
Hédl
R
Advancing the integration of history and ecology for conservation
Conservation Biology
 
2011
25
680
687
Thurstan
RH
Campbell
AB
Pandolfi
JM
Nineteenth-century narratives reveal historic catch rates for Australian snapper (Pagrus auratus)
Fish and Fisheries
 
2014
doi:10.1111/faf.12103
Van Houtan
KS
McClenachan
L
Kittinger
JN
Seafood menus reflect long-term ocean changes
Frontiers in Ecology and the Environment
 
2013
11
289
290
Walter
RC
Merrits
DJ
Natural streams and the legacy of water-powered mills
Science
 
2008
319
299
304
Willis
KJ
Birks
HJB
What is natural? The need for a long-term perspective in biodiversity conservation
Science
 
2006
314
1261
1265
Zu Ermgassen
PSE
et al
Historical ecology with real numbers: Past and present extent and biomass of an imperilled estuarine habitat
Proceedings of the Royal Society B
 
2012
279
3393
3400