Modern Earth system instrumentation and recordkeeping began in the late 1800s, with the Metre Convention in Paris on May 20, 1875 that established a standardized set of international weights and measures (Ortiz and Jackson, 2022). The 1700s and 1800s were crucial for European naturalists and emerging paleoclimatologists as they characterized the ice ages, evolution, and speciation (Stone, 1913; Hankins, 2015). Instruments to directly measure atmospheric gases, plant respiration, light refraction, and many other dynamics began to emerge, as scientists tracked precipitation, temperature, circulation, vegetation ranges, and animal movement utilizing standardized methodology (Edwards, 2011).

Today, as forecasting the range of potential climate extremes becomes critical, scientists must understand both the baselines and boundaries of physical and ecological dynamics. In many cases, long-term historical data is required as inputs to predictive models for forecasting climate change. For example, over 300 years of regional ecological data may be required to characterize critical global Earth system tipping points (Lenton et al., 2019), and ongoing system dynamics (Lenton et al., 2024; Miner et al., 2024), informing models on topics as diverse as regional temperature trends, sea level rise impacts, and glacier retreat timing. To forecast these physical dynamics, baselining data that predate the 19th-century scientific instrumentation revolution are required.

To date, the scientific community has often relied on global paleoclimate proxies including ice cores, tree rings, and isotopes as historical inputs for climate and ecosystem models (Winski et al., 2017; Christensen, 1993; Putnam et al., 2016). While these paleoclimate datasets store critical information, multi-decadal trends are often the easiest to derive. For example, while glacial ice cores may provide information about a particular glacier in addition to larger regional snowfall and temperature trends, larger-scale trends over multiple years in the global climate are most pronounced. Many proxies do not provide seasonal data, particularly for specific regions of interest, and there are substantial complexities in preserving and analyzing the records (Putnam et al., 2016; Pelletier et al., 2019; Birkel et al., 2011; Gajurel et al., 2020; Huhtamaa and Helama, 2017; Ungar et al., 2021).

Further, while the diversity of data available from modern satellite, airborne, drone, and in-situ measurements continues to grow (Bartsch et al., 2023; Schimel and Schneider, 2019), these data cannot capture changes predating the technological instrument expansion of the 1950’s (Miner et al., 2023). This dearth of information for baselining Earth system dynamics past the mid 1900’s increases uncertainty in forecasting the impacts of climate change.

As ecosystem baselines shift on even regional and seasonal scales (Lenton et al., 2019; Turner et al., 2020), there is increased urgency to fill the gaps between modern records and lower-resolution paleoclimate data. Expanding Earth system records with non-traditional data that spans centuries could support a greater understanding of current environmental change (Turner et al., 2020; Keeley, 2002). Therefore, data that can supplement and extend the instrument records past the last few centuries are necessary to support climate change planning.

Though not typically regarded as scientific data, non-traditional climate data including oral histories, personal written accounts, and art may inadvertently or directly record key aspects of long-term climate dynamics. From sea ice boundary observations to the seasonality of vernal flowering, pre-industrial lore and art documented weather, climate, and animal movement. Records of seafarers, land trade routes, crop surpluses and failures, and fishing hauls proliferate across cultures (Table 1; Figure 1). Though our understanding of the drivers of natural change has evolved, important data may underlie oral tradition and historical records.

While these records are often considered more a part of cultural than scientific tradition, folk or non-traditional data that predate modern scientific methods may be able to extend the current climate and ecosystem records through specific and careful application (Canavera, 2021; Bartlett et al., 2012; Canavera, 2023). For example, while traditional climate proxies tell us that our current rate of atmospheric warming is the most rapid in planetary history (Winski et al., 2017; Santibáñez et al., 2018), could non-traditional data help us validate or falsify assumptions, point us toward additional key research areas, or fill knowledge gaps?

Observations from British ship logs have already filled gaps in the Indian Ocean record for sea surface temperatures (Ingold, 2010). Style and color changes in Monet and Turner’s paintings have been used to infer trends in air pollution during the Industrial Revolution (Martin and Barboza, 2020). Glacier landscapes in paintings have been used to extend photographic records of long-term glacier retreat (Lacina and Halas, 2015; Winiger et al., 2019). Despite these successes, non-traditional climate data have not been applied to many critical climate change questions due to the challenges in confirming the authenticity and accuracy of the records. For example, it may be difficult to ascertain the influence of key artistic choices or elements driven by imagination or cultural expectations. As a result, the applied uses of oral, written and artistic records may be often overlooked.

This review is a first effort to identify and catalog a wide range of non-traditional climate and ecosystem data from diverse sources that could be applied to extend the instrument record or pre-screen modern remote sensing retrievals. The compiled records span fine art (Lacina and Halas, 2015; McLoughlin, 1999; Farag-Miller et al., 2013; Metzger, 2020; Sillasoo, 2014; Rees, 2008; Rudolph, 1966; Ingold, 2010), shipping manifests (Christensen, 1993; D’Alto, 2007; García-Herrera et al., 2018; Catchpole and Faurer, 1985; Teillet, 1988), oral and written compendia, species movement records (Stone, 1913; Stejneger, 1905; Arnott, 2007; Linglin et al., 2020; Reeb et al., 2020; Martin et al., 2018; Martin and Barboza, 2020), and trade reports (Sahrhage and Lundbeck, 1992; Wilcox, 1982; Jurabek, 2022). While not geographically or temporally comprehensive, the non-traditional data reviewed here offer a perspective of the possible climate applications in specific regions. We highlight these records as an example of diverse non-traditional data that could be useful in establishing regional ecological and climate baselines that predate instrument records (Miner et al., 2023).

While each non-traditional dataset will need to be assessed individually for utility and cross-checked for accuracy, any opportunity to extend the climate record cannot be overlooked. If historical non-traditional data sources can provide an additional index for climate change, the observations and stories of diverse naturalists, healers, artists, historians and explorers across the globe could supplement modern observations.

To forecast the impacts of climate change, we must fully understand the recent and distant history of physical and ecosystem processes. Reconstructing the climate signal at spatiotemporal scales between the paleo record and modern instrument science has been an ongoing challenge, increasing uncertainty. It is possible that with the careful application of non-traditional data records, key information could be extracted to reconstruct past climate and provide context for contemporary observations.

To successfully apply these non-traditional data, it will be critical to develop methodologies for standardizing and scaling records across regions and observers. For example, using industrial fish haul records together with seasonal fishing lore could help illuminate the ecosystem dynamics of a specific place and time. In some cases, these records may reflect the cultural traditions or colonialism of a specific time and must be taken within the context of a greater historical perspective. In this regard, pre-colonial written or oral records from many cultures could be compiled through additional effort within specific local archives.

A methodology incorporating cross-verification could strengthen the applicability of non-traditional data and help identify patterns of bias or subjectivity, using stories and measurements from one discipline to understand the veracity of another. To accomplish this, diverse non-traditional data from the same region could be applied to research questions, informing the context and characterizing data gaps. Similar normalizing techniques for standardizing non-traditional data have been used for specific use cases (Martin and Barboza, 2020; Robinson, 2005; Albright and Huybers, 2023; Erlandson and Rick, 2010) and could be applied more broadly.

The use of artificial intelligence and image extraction tools would also be of specific benefit when characterizing both the availability and content of non-traditional data. Applying AI tools would allow the extraction of content from paintings and records recorded as copies or images. While citizen science projects have successfully implemented volunteers to identify difficult to read data including sea records of monsoons and fish hauls (Zooniverse, n.d.), identifying gaps and trends across the resulting large datasets is an ideal application for AI (Gay et al., 2023). As AI becomes increasingly useful for interpreting large datasets, the most modern tools help us understand records from the distant past.

Folklore endures. Though records have been lost to fires, plagues, violence, and colonization, many stories passed down through generations persist. Non-traditional data could fill gaps in understanding biodiversity and landcover change, provide a longer record to strengthen model projections, and baseline changing ecosystems (Lenton et al., 2019; Heinze et al., 2021). In this time of unprecedented environmental change, it is more important than ever to elevate and utilize all the ways of knowing and understanding the Earth, allowing the distant past to help us prepare for the future.