This article is protected by copyright. All rights reserved.
(e.g., Hinckley et al., 2016a). However, progress toward this goal is limited by lack of understanding of ecological insights that can be gained through syntheses of existing data, including testing of outstanding hypotheses and the generation of new hypotheses (LaDeau et al., 2017). Specifically, there is a lack of understanding of how the complementary structures of various networks might be used to formulate research syntheses. In addition, research agendas or frameworks are lacking that connect research questions to available data for specific combinations of existing networks in ecology and environmental science.
This paper aims to fill these gaps. We explore the potential for combining long-term experimental results and hypotheses from research networks with highly standardized long-term observations from monitoring networks to elucidate the mechanisms that drive long-term ecological and environmental change. Our objectives are to:
1) Describe types of environmental science networks and their complementary features 2) Assess the progress to date for cross-network synthesis studies of LTER and NEON, and 3) Identify opportunities and challenges that build on the work accomplished to date.
We highlight potential synergies between the Long-term Ecological Research (LTER)
Program, a research network, and the National Ecological Observatory Network (NEON), a monitoring network, both funded by the US National Science Foundation (NSF) (Collins & Childers, 2014). Both networks address major challenges in environmental science and make their data publicly available for use by researchers, educators, policy-makers and others. Our findings are also relevant to other research and monitoring networks in the United States and internationally (Richter et al. 2018). These networks include the Critical Zone Observatories (CZO) (White et al., 2015) funded by NSF; the Forest Service Experimental Forests and Ranges (e.g., Lugo et al., 2006) and agricultural experimental watersheds and ranges (Bartuska et al.,
This article is protected by copyright. All rights reserved. 2012) funded by the U.S. Department of Agriculture; the AmeriFlux network funded by the Department of Energy (Novick et al., 2018); the international Global Lakes Ecological Observatory (GLEON) (Hanson et al., 2016);programs managed by the United States Geological Survey (USGS, 2016); and the cooperative National Atmospheric Deposition Program (NADP) (see Supporting Information).
In this paper, we describe the results of an NCEAS working group on LTER-NEON synergies. The working group included scientists from LTER, NEON, and the broader ecological community whose research draws on environmental research networks. In two workshops and successive discussions, we analyzed the structure of LTER and NEON and their complementarities (Section 2), created a typology of synthesis efforts (Section 3), and evaluated the progress to date and challenges and opportunities for future efforts in six broad research areas of ecosystem and environmental science (Section 4).
While many networks encompass both aspects, research and monitoring networks have distinct designs and administration (Table 1, Table S1, Figure S1). Research networks (e.g., LTER, CZO, GLEON, US Forest Service Experimental Forests and Ranges [USFS EFR]) focus on question-driven research, based on observational studies and experiments that test mechanistic hypotheses about ecological processes, and are designed and conducted by a community of researchers, who make their data available. Sites may be funded individually, and may seek renewed funding on a competitive basis, based on agency guidelines and priorities.
Sites in research networks may adopt and extend prior long-term studies, and engage in synthesis efforts across sites, but synthesis among sites may be limited by inconsistent methods. In
This article is protected by copyright. All rights reserved. contrast, monitoring networks (e.g., NEON, NADP, USGS National Water Information System
[NWIS]) focus on long-term, standardized data collection of patterns in a set of pre-determined variables, based on a pre-defined sampling design. Science and technical staff manage instruments, lab analyses, and data collection, quality control, and archiving procedures. Data collection involves standardized protocols, sensors, and technologies and data are collected using a pre-defined sampling frequency. Sites were selected and are funded as a group, for a specified period. While the dichotomy of network types illustrated in Table 1 represents well the differences between LTER and NEON, many other networks share features of both research and monitoring networks as defined here.
We identify synergies between research and monitoring networks, using the US LTER Program as an example research network and the US National Ecological Observatory Network (NEON) as an example monitoring network (Figure 1). Synergies between LTER and NEON arise from their complementary designs: LTER focuses on mechanistic understanding of ecological processes, and provides conceptual models, hypothesis-testing, long-term experiments, temporal coverage and information management, while NEON focuses on quantification of ecological trends, and provides consistent design, standardized measurements, spatial coverage, and a data resource (Figure S1).
The LTER Program, a research network, was initiated in 1980 and presently includes 28 sites in a wide range of ecosystems (Callahan 1984, https://lternet.edu/site/, Figure 2, Table S1, Table S2). Researchers propose sites, establish the research agenda at each site, and conduct research on long-term ecological processes. The five core areas: primary production, population studies, movement of organic matter, movement of inorganic matter, and disturbance (https://lternet.edu/core-research-areas/) provide a research framework for synergies that address
This article is protected by copyright. All rights reserved. major questions in environmental science (Figure 1). Data are available from the Environmental Data Initiative (https://environmentaldatainitiative.org/). NEON is a monitoring network, which began to provide data in 2015, and was designed to examine ecological change over time at a set of 47 terrestrial and 34 aquatic sites selected to represent the diversity of eco-climatic domains in the continental U.S. (Kampe et al., 2010;Kao et al., 2012;Goodman et al., 2014;Springer et al., 2016;Thorpe et al., 2016; https://www.neonscience.org, Figure 2, Table S1, Table S2). The network includes 30-year installations in core 'wildland' ecosystems within each of the 20 NEON domains as well as additional sites that cover environmental variability within the domain. The major measurements of NEON include flux tower measurements, airborne remote sensing, aquatic measurements, soil measurements, and terrestrial organism sampling to document how U.S. ecosystems are changing (Figure 1, Supporting Information). NEON data are available via the NEON data portal (https://www.neonscience.org/data/about-data/getting-started-neon-data).
NEON was designed by the research community, including LTER researchers. Its topdown standardized measrurment programs complement the bottom-up, research-question driven approaches in LTER. Proximity to LTER sites was one factor considered in selecting NEON sites. Hence, cross-network syntheses with LTER were envisioned from the beginning of NEON.
Both LTER and NEON address issues of broad social relevance. Social science is not a core area of LTER (Jones and Nelson, 2020), but social-ecological systems and related questions (e.g., Collins et al., 2011) are central to many LTER programs. Consideration of socialecological systems motivated the selection of many of the variables measured by NEON, such as disease-transmitting organisms, that are not consistently measured at LTER sites. Social and economic factors also are central to international LTERs (Mirtl et al., 2018).
This article is protected by copyright. All rights reserved.
Cross-network synthesis efforts are needed, and can be very powerful, because of the insights they provide. Goals for such syntheses include: generalize patterns and processes among locations, identify interactions among ecological processes at one or multiple sites, generalize across temporal scales, reveal differences among methods, or test the potential and limitations of models (Table 2). One form of synthesis tests the generality of a finding (or concept or hypothesis) about a single property or process across sites (Type 1, Table 2). Examples include how C flux varies among locations in the AmeriFlux network (Novick et al., 2018); how atmospheric deposition varies among locations in the NADP network (Lajtha & Jones 2013); how streamflow trends vary among locations in the USGS NWIS (Lins & Slack 1999); or how climate trends vary among locations in the US Historical Climatology Network (USHCN) (Menne et al., 2018). Type 1 synthesis spans the geographic coverage of the networks. For example, the 28 LTER sites and the 47 NEON sites are distributed throughout the United States and LTER sites also occur in Antarctica and the Pacific (Figure 2).
Analysis of multiple data streams can produce a more complete or nuanced understanding of a phenomenon or opportunities to test hypotheses using independent datasets. A second type of synthesis, "multiple properties or processes within a site," aims to elucidate interactions among ecological processes using data on multiple complementary properties or processes at a site (Type 2, Table 2). Long-term mechanistic experiments (e.g., from LTER) provide insights for interpreting monitoring data (e.g., from NEON) at a site. Examples include how long-term manipulations of vegetation influence C exchange, or how an invasive insect affects ecosystem water exchange (Giasson et al., 2013;Kim et al. 2017). Type 2 synthesis studies could be based
This article is protected by copyright. All rights reserved. on complementary measurements from multiple networks. Many opportunities for Type 2 synthesis exist at sites which are "co-located" (participate in) both the LTER and NEON networks (Figure 2, Table S1).
Analyses of multiple data streams from different networks could contribute to more general understanding of patterns and trends at regional to continental scales over the long term.
A third approach to synthesis, "multiple properties or processes across sites" (Type 3, Table 2) seeks generalizations about interactions among ecological processes at many locations. For example, long-term experiments at multiple locations provide insights for interpreting monitoring data within or among biomes or ecosystem types, such as how vegetation manipulations affect streamflow in multiple different forest ecosystems (Jones & Post, 2004), or how climate change is affecting ecosystem water use (Jones et al., 2012). Type 3 synthesis studies could be based on complementary measurements from the nine co-located sites in LTER and NEON (italicized in Table S1, Figure 2, Figure 3a), grouped by biome or ecosystem type, or across all sites in the two networks, which span much of the range of mean annual precipitation and temperature in North America (Figure 3b) (Villarreal et al., 2018).
Additional forms of synthesis among research and monitoring networks include syntheses across scales, across methodological approaches, and using modeling (Table 2). Syntheses across scales (Type 4) build on data collected at more than one temporal scale to elucidate temporal patterns in ecological processes, including trends, cycles, and thresholds. For example, long-term datasets and experiments at research networks such as LTER complement short-term, high-resolution data from NEON or other monitoring networks. Syntheses that compare methods (Type 5) can reveal differences in ecological patterns that result from disparate measurement approaches. Different types of data pertaining to a single phenomenon permit
This article is protected by copyright. All rights reserved. comparisons among multiple modes of observation. Model syntheses (Type 6) combine data from mechanistic experiments and monitoring to inform and constrain models, to understand uncertainty in projections, and to identify needs for model improvements.
We provide examples of potential cross-network synthesis studies that can accelerate environmental science by linking the five core areas of long-term research (in LTER) with the five main measurement programs (of NEON) (Figure 1). These examples comprise six broad research areas: 1) ecosystem fluxes of C and energy; 2) remote sensing and ecosystem models;
3) aquatic-terrestrial linkages; 4) soil biogeochemical and microbial dynamics; 5) organism and species distribution models; and 6) land use and disturbance history, resilience and stability (Figure 1). A key theme in all these examples is how synergies emerge from the interaction of LTER hypothesis-based, mechanistic science interacting with NEON standardized, spatiallydistributed monitoring.
Complementarities among research and monitoring networks can address broad questions about C cycling at regional and continental scales (Figure 1). Long-term experiments and observational studies have documented multi-decade changes in ecosystem C storage, and the mechanisms underlying these changes. These experiments and studies complement highresolution information on C exchange from eddy flux towers. In the Arctic tundra and boreal forest (LTER sites in Alaska), warming climate has reduced soil C stocks (Euskirchen et al., 2017). In the desert (Sevilleta LTER), vegetation change from grassland to shrubland increased C sequestration (Petrie et al., 2015). In freshwater marsh and mangrove forests (Florida
This article is protected by copyright. All rights reserved.
Everglades LTER), C sequestration depended on vegetation type, temperature, and flooding (Malone et al., 2016). In a temperate freshwater marsh (Plum Island LTER), increased rainfall reduced soil salinity thereby increasing productivity and C (Forbrich et al., 2018). In an urban site (Phoenix LTER), outdoor water use increased evapotranspiration as well as C storage (Templeton et al., 2018). There is great potential synergy between these experiments and networks of eddy covariance flux towers that provide continuous measurements of carbon dioxide (CO2), water vapor, and energy fluxes that are used to estimate ecosystem productivity and water and C exchange between ecosystems and the atmosphere (Campioli et al., 2016). The global network of eddy flux towers (e.g., AmeriFlux, Fluxnet) (Novick et al., 2018) 3). For example, in northern temperate forest (the Harvard Forest LTER and NEON site), combining data from long-term experiments on vegetation manipulation from LTER with information from multiple eddy flux towers (Type 2 synthesis) revealed how soil respiration varied with weather, phenology, invasive insects, forest management practices, and atmospheric N deposition over 22 years (Giasson et al., 2013). In another example, longterm monitoring of vegetation and streamflow from LTER was combined with data from flux towers to show how invasive insect outbreaks reduced leaf area and increased water yield at the Harvard Forest (Kim et al., 2017).
This article is protected by copyright. All rights reserved.
Going forward, there is great potential for testing general hypotheses by combining information from LTER long-term experiments or observations with data from multiple eddy flux towers maintained by NEON, LTER, or other networks, which replicate measurement conditions or sample local variation in vegetation or landform conditions (Table 3). Examples include how past fire severity affects C flux in the Arctic tundra (Rocha & Shaver 2011); how various permafrost conditions affect C flux response to climate warming in the boreal forest (Bonanza Creek) (Euskirchen et al., 2014); and how various land use histories or insect invasions affect C exchange (Harvard Forest, Figure 4). Moreover, such studies could show how local variation in soil temperature, water table fluctuations, and plant activity (measured by LTER and NEON) affect C flux (measured at eddy flux towers in NEON, LTER, and other networks) (e.g., Sturtevant et al., 2016). Combining information from multiple towers at a site can assist efforts to scale up eddy fluxes for modeling (Xu et al., 2017) (Table 3, Table S3).
In the future, Type 3 (multiple properties or processes across sites) syntheses could contribute to more general understanding of C and water exchange over the long term (Figure 1). For example, Type 3 syntheses could combine long-term observations of vegetation and climate at LTER sites with data from eddy flux towers to test hypotheses about how trends in winter precipitation influence C uptake in warm desert shrublands (e.g., Biederman et al., 2018). Type 6 (models) syntheses could combine results from long-term experiments and observations at LTER sites with data on C exchange from eddy covariance sites from NEON, LTER, and AmeriFlux in order to test hypotheses linking rising atmospheric CO2, plant functional traits and forest structure, and ecosystem water use efficiency in forests (e.g., Mastrotheodoros et al., 2017).
Type 4 (across temporal scales) and Type 6 (models) syntheses also could combine data from long-term experiments and monitoring with shorter-term eddy flux data in models to predict the
This article is protected by copyright. All rights reserved.
response of net ecosystem exchange to long-term ecosystem change (e.g., Wright & Rocha 2018).
Several forms of synthesis could combine long-term field data from LTER with lidar and hyperspectral data from NEON to assess how land cover change and vegetation dynamics influence ecosystem processes (Figure 1). Each year, the NEON Airborne Observing Platform obtains acquires lidar (Light Detection and Ranging) and imaging spectrometer data with a nominal spatial resolution of 1-2 m 2 , and 0.25 m resolution digital orthophotos for hundreds of square kilometers surrounding each NEON site (Figure 2). Data are made available at various post-processing levels and include topography, vegetation structure, and canopy physical and chemical properties.
Initial efforts have used field data in combination with NEON airborne mapping products to improve remote-sensing based vegetation classifications (e.g., Scholl et al., 2020), infer structures that may influence ecosystem function (LaRue et al., 2019), or to map biodiversity patterns that are difficult to assess from field data (Hakkenberg et al., 2018;Musavi et al., 2017).
Type 2 (multiple properties or processes within sites) and Type 3 ((multiple properties or processes across sites) syntheses could combine field data from LTER with NEON remotelysensed data to explore how landforms influence disturbance, climate, and vegetation dynamics (e.g., Antonarakis et al., 2014;Frey et al., 2016;Yousefi Lalimi et al., 2017). Repeat NEON mapping using hyperspectral imagery may reveal ecosystem responses, such as plant water stress (e.g., Brodrick & Asner 2017), that correspond with long-term trends in vegetation measurements from LTER. Type 2 or type 3 syntheses also could combine long-term data on
This article is protected by copyright. All rights reserved. vegetation from LTER sites with analyses of NEON's laser scanning and imaging spectroscopy to examine how ecosystem changes are related to plant functional traits such as foliage height diversity, leaf chlorophyll and water content (e.g., Schneider et al., 2017) or plant biomass (Goulden et al. 2017).
Type 6 syntheses (models) have used NEON eco-climate domain polygons as the basis for efforts to extrapolate ecosystem processes across regions. For example, Iwema et al. (2017) used data from the AmeriFlux network to examine how soil moisture measurements in eight
Combining long-term experiments and observations from LTER with data provided by NEON could improve whole-catchment C and nutrient budgets (Figure 1). Although they
This article is protected by copyright. All rights reserved. occupy small areas, aquatic ecosystems can make a disproportionately large contribution to terrestrial C storage in some regions (Buffam et al., 2011), and rivers export a significant fraction of terrestrial net ecosystem production in the U.S. each year (Butman et al., 2016). However, spatial variability of C storage and transport is high (Argerich et al., 2016) and strongly linked to terrestrial processes (McCullough et al., 2018), including past disturbance (Meyer et al., 2014, Lajtha & Jones 2018). Terrestrial ecosystem processes also influence spatial and temporal variation in N export from streams (Beaulieu et al., 2015, Neilson et al., 2018, Webster et al., 2019). Improved integration of ecosystem properties linking terrestrial and aquatic ecosystems would more accurately reflect C and nutrient budgets at scales relevant to Earth system models (Wollheim et al. 2018).
In the future, Type 3 (multiple properties or processes across sites) analysis of linked aquatic-terrestrial dynamics could link aquatic and terrestrial installations at NEON sites, some of of which are co-located with LTER (or other network) eddy flux tower (Table 4, Table S4).
loading may also influence N and P in lakes (Corman et al., 2018) or how N and P loading influence C dynamics in streams (Mutschlecner et al., 2018). They can have shown how fire and grazing influence inorganic nutrient dynamics of streams (Sullivan et al., 2019), or how ecosystems process N deposition (Litaor et al., 2018). Fluorescence measurements of dissolved organic matter have helped discriminate land use and climate effects on the chemistry of exported DOC at the Andrews Forest LTER in Oregon (e.g. Lee & Lajtha, 2016); fluorescent DOC measured at many NEON sites could be used in future syntheses linking multiple sites. At the North Temperate Lakes LTER in Wisconsin, the contribution of lakes to total CO2 flux can be estimated by combining LTER lake metabolism and CO2 data and models with terrestrial flux
This article is protected by copyright. All rights reserved.
estimates from a nearby NEON terrestrial flux tower (Table 3). Combined NEON and LTER installations will facilitate estimates of allochthonous and autochthonous sources of C in aquatic ecosystems (Hanson et al., 2016), which are essential to constructing C budgets at regional to continental scales.
Integration of NEON with LTER and other networks might also advance understanding and prediction of N fluxes at continental scales. For example, Type 3 studies could use NEON and LTER data to test how vegetation cover and phenology from remotely sensed imagery are related to stream N fluxes in various biomes (Table 4). Trends and fluxes of N in precipitation or streams that have been described for various networks (e.g., Argerich et al. 2013, Lajtha & Jones, 2013) could be combined in Type 3 (multiple properties or processes across sites) studies with aquatic N data from NEON aquatic sites and N content of plant canopies in those watersheds, estimated from hyperspectral data collected by the NEON Airborne Observing Platform.
Many opportunities exist to combine long-term studies from LTER with soil measurements from NEON to better understand how soil biogeochemistry and microbial processes drive ecosystem response to environmental change (Figure 1). Long-term experiments on soil N additions, soil warming, and soil detrital additions and removals have been conducted at many locations, including LTER sites. NEON samples biogeochemical stocks and soil N processes, microbial community composition and biomass, and vegetation one to three times per year at five-year intervals at multiple plots in each NEON site (Hinckley et al., 2016b).
This article is protected by copyright. All rights reserved.
Type 2 (multiple properties or processes within sites) and Type 3 (multiple properties or processes across sites) synthesis efforts could inform our understanding of drivers of microbial abundance, diversity, and community composition; organic matter and nutrient cycling dynamics; and C stabilization and N transformations across different ecosystems. For example, long-term experiments have shown that despite strong compositional differences across sites, microbial communities shifted in a consistent manner in response to N or P additions (Leff et al., 2015) as well as climate variability and soil C content (Delgado-Baquerizo et al., 2016). Type 2 and Type 3 syntheses could link long-term experiments at LTER sites to data on microbial populations, climate, and nutrient fluxes from NEON sites to examine hypotheses about how microbial dynamics mediate biogeochemical fluxes.
In addition, new syntheses could improve predictions of which systems are most vulnerable to C and nutrient loss at regional to global scales. A long-term experiment showed that two decades of elevated nitrogen inputs increased forest soil C, largely due to a suppression of organic matter decomposition (Frey et al., 2014). Type 3 syntheses could link the findings from long-term soil nutrient addition experiments and soil warming experiments to soil surveys and distributed NEON data (e.g., soil C and N concentrations and stocks) to predict soil C and N sinks and sources at the continental to global scales (e.g., Crowther et al., 2016;Wieder et al., 2015).
LTER studies also have shown that soil C and N responses to long-term warming and nutrient additions vary seasonally (Contosta et al., 2011) and may continue to change over multiple decades (Melillo et al., 2017;Reich et al., 2018). Type 2 syntheses could enhance understanding of soil N response to environmental change by combining long-term experiments and observations of effects of atmospheric N deposition, windthrow, fire, grazing, and other
This article is protected by copyright. All rights reserved. changes at LTER sites with NEON observations of soil properties at those sites (Figure 1). Type 6 syntheses (models) could use ecosystem models that combine long-term data on atmospheric deposition from NADP (Lajtha & Jones 2013;Sullivan et al., 2018) with NEON's standardized N mineralization data to predict and interpret effects of air pollution on soil ecosystem processes.
Syntheses linking long-term observations and experiments from LTER with data from NEON can provide insights into population and species responses to environmental change (Figure 1). Long-term observational studies reveal how populations and communities respond to land use, disturbance, and climate. NEON provides data on microbial communities, aquatic and terrestrial plants, breeding birds, and fish, as well as focal species of small mammals and insects.
NEON is also analyzing eDNA (i.e., organism DNA in the environment) in aquatic ecosystems.
Environmental DNA has great potential for monitoring common species and to detect and identify the presence of many species (Bohmann et al., 2014).
Synthesis of new data from NEON with long-term studies can address key questions in biodiversity, population dynamics, species distribution models, and meta-community dynamics.
For instance, long-term studies at Harvard Forest LTER in Massachusetts have shown that small mammal community structure is relatively unaffected by species invasion (e.g., of hemlock woolly adelgid) or disturbance (e.g., experimental mortality of hemlock) (Degrassi, 2018).
Spatial analyses of small mammal data across the continental United States (from NEON) indicated that body size variation and mammal species richness were positively associated with temperature (Read et al., 2018). Type 2 or Type 3 synthesis efforts could combine results of long-term experiments from LTER showing mechanistic organism response to invasion or
This article is protected by copyright. All rights reserved. disturbance with data from NEON sampling of small mammals to examine how climate change, disturbance, and invasion processes are affecting mammal populations at local or continental scales. For example, at the Konza Prairie in Kansas, NEON small mammal and tick survey plots are located in areas where fire and grazing have been manipulated in LTER long-term experiments, potentially revealing how small mammal and tick populations respond to disturbances (Figure 5). The co-location of LTER experiments and NEON sampling enable NEON data to reveal novel results of LTER long-term experiments, while concurrent LTER data collection and complementary experiments provide mechanistic explanations and context for interpreting species data from monitoring networks.
Type 2 (multiple properties or processes within sites) and Type 3 (multiple properties or processes across sites) syntheses also can unravel underlying causal mechanisms linking longterm fish population responses to environmental change by combining systematic fish surveys and eDNA measurements in aquatic systems conducted by NEON to results from long-term experiments and observations. Long-term studies at LTER sites have documented native fish population responses to invasive fish species (Hansen et al., 2017), to climate and trophic interactions (Parks & Rypel, 2018), and to disturbance and vegetation change (Dodds et al., 2012). Initial studies indicate that eDNA can be used in conjunction with long-term monitoring of fish populations in lakes (Klobucar et al., 2017). Given the high variance of many aquatic populations over time (e.g., Batt et al., 2018), Type 2 (multiple properties or processes at a site)
or Type 6 (model) syntheses that combine NEON data on both fish population dynamics and physicochemical conditions within lakes and streams with LTER and other long-term studies of streams and lakes will help reduce uncertainty in population models and causes of population change in fish.
This article is protected by copyright. All rights reserved.
Several forms of synthesis efforts also could contribute to species distribution models (Figure 1). NEON is collecting systematic data on focal taxa, including soil microbes; ticks, mosquitoes, and ground beetles; small mammals; and breeding birds (Springer et al., 2016;Thorpe et al., 2016;Egli et al., 2020). LTER studies have documented long-term trends and constructed models for species distributions of birds (Betts et al., 2018), arthropods (Lister & Garcia, 2018), and invasive insects (Schliep et al., 2018). Long-term experiments and observational studies also document community level responses, for example, to species loss (e.g., hemlock removal, Record et al., 2018) or disturbance (e.g., saltwater intrusion, Zhai et al., 2016). NEON data are being used to model spatial patterns of tick abundance (Klarenberg & Wisely 2019). Data from long-term experiments and observations have been used to test ecological theory and improve models of species distribution and dynamics (e.g., Thomas Clark et al., 2018. Snell Taylor et al., 2018). Going forward, synthesis studies could combine results from long-term experiments at LTER sites with measurements of species across the NEON network to draw inferences about general factors influencing species distributions (Type 3) and to identify knowledge gaps in models of species distributions (Type 6).
Combining theory, long-term experiments, and observations of disturbances from LTER with high-frequency data provided by NEON can provide insights and generalizations about ecosystem response to disturbance, stability and resilience (Figure 1). Long-term studies demonstrate how land use legacies and disturbance history shape modern-day landscape patterns and ecological communities (Acker et al., 2017). Data from NEON, including flux towers, remote sensing, aquatic, soil, and organism sampling, provide information on current ecosystem
This article is protected by copyright. All rights reserved.
status. The combination of long-term experiments and observations from LTER on land use history and disturbance could provide context for understanding monitoring data from NEON on ecosystem fluxes (Figure 4) and species distributions and abundance (Figure 5).
Long-term studies have documented alternative stable states and associated mechanisms, but records may be insufficient to test tor regime shifts (Bestelmeyer et al., 2011, Ratajczak et al., 2014, Yu et al., 2019), because detection of their approach and validation of the change in feedbacks that accompany regime shifts require unbroken series of frequent observations sustained for long periods of time (Butitta et al., 2017). LTER-developed theory (e.g., Adam et al., 2011;Bestelmeyer et al., 2013;Chapin et al., 2010) provides a framework for combining long-term data from LTER with high-resolution NEON data to gain insight into ecosystem resilience. In aquatic systems, NEON will collect high-frequency (sub-hourly) measurements of several variables including nitrate, dissolved organic matter, and conductivity. Type 4 syntheses (across temporal scales) of NEON data combined with LTER data and understanding of ecosystem states could help test a key hypothesis that changes in the variance of biogeochemical fluxes (P, N, and C) may reveal regime shifts in ecosystems (e.g., Webster et al., 2016) (Figure 1). In summary, many examples exist of ongoing synthesis between the LTER research network and the NEON monitoring network, but these are mostly Type 2 (multiple properties or processes within sites) syntheses based at sites that are co-located in both networks. While a great many papers have been published describing the potential for LTER-NEON syntheses, very few studies have been published that report results of such syntheses. Moreover, there is a dearth of studies that utilize the many other potential types of synthesis, including Type 3 (multiple properties or processes across sites), Type 4 (across temporal scales), Type 5 (across
This article is protected by copyright. All rights reserved. methodological approaches), and Type 6 (models) syntheses. Nevertheless, as described above, ongoing work provides exciting potentials for specific research questions that could be explored using these varied synthesis approaches.
In this era of rapid, broad-scale environmental change, publicly available information from complementary environmental science research networks, such as LTER, and monitoring networks such as NEON offer opportunities for discovery, arising from the potentials for LTER measurements, experiments, models, and observational studies to provide context and mechanisms for interpreting NEON data, and for NEON measurements to provide standardization and broad scale coverage that complement LTER studies. Many different types of cross-network synthesis are possible, in six broad areas of ecology. To date, cross-network efforts are addressing topics including how long-term vegetation change influences C fluxes; vegetation structure and function revealed by detailed remote sensing; aquatic-terrestrial connections of nutrient cycling linking vegetation to streams and lakes; effects of soil biogeochemistry and microbial processes on ecosystem response; population and species responses to environmental change; and ecosystem response to disturbance, stability and resilience. Current efforts focus primarily on synthesis of properties and processes at individual sites where NEON and LTER are co-located, but they could be extended in ways described in this paper to address broader questions in ecology and environmental science, at a wider range of sites. These potential syntheses also provide a pathway for the broader scientific community, beyond LTER and NEON, to participate in cross-network research. These findings apply to cross-network syntheses among other research and monitoring networks in the US and globally,
This article is protected by copyright. All rights reserved. and can guide scientists and research administrators in promoting broad-scale research that supports resource management and environmental policy. The emergence of these synergies should also help to make long term research networks and sites more open to new investigators as they will facilitate the flow of information and ideas and the development of new collaborations. This flow, and the links to resource management and policy could also contribute to broadening participation of groups traditionally under-represented in science.
Table 4. Examples of potential future synthesis opportunities to characterize whole-watershed elemental budgets and examine interactions between aquatic and terrestrial ecosystem processes by combining aquatic studies, eddy flux tower, and other measurements at sites that are co-located in LTER (research network) and NEON (monitoring network). See details of site locations and instrumentation in Table S4. LTER sites marked with asterisk are located in the same watershed as the NEON sites.
LTER site name NEON Aquatic site NEON terrestrial site Arctic tundra Arctic (ARC) Oksrukuyik Creek (OKSR) Toolik (TOOL) Temperate forest Baltimore (BES) Posey Creek (POSE) Blandy Experimental Farm (BLAN) Boreal forest Bonanza (BNZ)* Caribou Creek (CARI) Caribou-Poker Flats watershed (BONA) Savanna Georgia Coastal Ecosystem (GCE) Barco Lake, Suggs Lake (BARC, SUGG) Ordway-Swisher Biological Station (OSBS) Grassland Konza (KNZ)* Kings Creek (KING) Konza Prairie Biological Station (KONZ) Alpine tundra Niwot Ridge (NWT) Como Creek (COMO) Niwot Ridge (NIWO)
This article is protected by copyright. All rights reserved. they link research concepts to monitoring data. In this example, core areas of inquiry in a research network (LTER) and major areas of standardized measurements in a monitoring network (NEON) provide complementary contributions to potential synergies that address key questions in environmental science. shown as open symbols; LTER sites are shown as closed symbols. Numbers refer to NEON ecoclimatic domains, and three-letter acronyms refer to LTER sites (see Table S1). Further details on site locations are available at https://lternet.edu/site/ (LTER) and https://www.neonscience.org/about-neon-field-sites (NEON). temperature (MAT) and precipitation (MAP) for LTER sites and core terrestrial NEON sites, oriented in Whittaker biome space. A total of nine LTER sites are co-located with NEON sites: seven (ARC, BNZ, HFR, KNZ, NTL, NWT, SGS) are co-located with core terrestrial NEON sites, and two (AND, JRN) are co-located with non-core NEON sites. The three-letter acronyms for LTER sites and the D01 notation for NEON domains are defined in Table S1. Climate data from co-located sites are enclosed within circles. Co-located sites may have slightly different climate values because climate varies within each domain, and climate data may have been obtained from different meteorological stations and/or for different time periods (see Table S1).
. . . . . . . . . . . . . . . . ! . 0 875 1,750 2,625 437.5 Kilometers 0 210 105 Kilometers 0 230 115 Kilometers 0 920 1,840 460 Kilometers . NEON ! LTER ! ! ! ! . . -25 -20 -15 -10 -5 0 5 10 15 20 25 30 0 500 1000 1500 2000 2500 3000 3500 4000 MAT (°C) MAP (mm) NEON core terrestrial LTER terrestrial and coastal MCM ARC-D18 BNZ-D19 NWT-D13 NTL-D05 HFR-D01 KNZ-D06 LUQ D09 D12 SGS-D10 D15 AND-D16 D20 CWT D08 BES D02 D07 D11 HBR SBC JRN-D14 SEV D17 CAP KBS PIE GCE D03 FCE D04 Tundra Temperate forest Tropical dry forest Tropical wet forest Boreal forest Desert Savanna -25 -20 -15 -10 -5 0 5 10 15 20 25 30 0 500 1000 1500 2000 2500 3000 3500 4000 MAT (°C) MAP (mm) LTER terrestrial and coastal NEON