Critical Zone Research and Observatories: Current Status and Future Perspectives

The Earth’s Critical Zone (CZ) is defined as the thin layer of the Earth’s surface and near-surface terrestrial environment from the top of the vegetation canopy (or atmosphere–vegetation interface) to the bottom of the weathering zone (or freshwater–bedrock interface) (National Research Council, 2001). This zone encompasses the near-surface biosphere, the entire pedosphere, the surface and near-surface portion of the hydrosphere and the atmosphere, and the shallow lithosphere (Lin, 2010). This concept of the CZ provides a unifying framework for integrating belowground–aboveground, abiotic–biotic, and time–space in mass and energy flows to holistically understand complex terrestrial ecosystems and offers a fertile ground for interdisciplinary research (Anderson et al., 2007; Lin et al., 2011). Thus, the integrated study of the CZ has been recognized as one of the most compelling research fields in Earth and environmental sciences in the 21st century (National Research Council, 2001, 2011).

Environmental processes within the CZ, such as mass and energy exchange, soil formation, streamflow generation, and landscape evolution are crucial to sustaining biodiversity and humanity (Lin et al., 2011; Field et al., 2015). The CZ supplies nearly every life-sustaining resource on which life originates, evolves, and thrives (National Research Council, 2001; Lin, 2014). This zone provides diverse services to human society and determines human livelihood (Lin, 2014; Field et al., 2015). Knowledge of how the CZ forms, functions, and supports humanity is an increasingly important issue raised by both the general public and the scientific community (Banwart et al., 2013). However, with accelerated socioeconomic development, the CZ is under ever-increasing pressure from human perturbations, such as the rapid growth of human and livestock populations, land use intensification, global environmental changes, and expanding consumption patterns (IPCC, 2013). The rapidly expanding needs for sustainable development give a special urgency to better understand, predict, and manage the complexity and dynamics within the CZ and its interactions with other environmental systems (Lin, 2010; Banwart, 2011).

Investigating and understanding the CZ requires a synergistic approach across disciplines, including soil science, hydrology, biology, ecology, geology, geomorphology, geochemistry, geophysics, geobiology, and many others (Anderson et al., 2008; Brantley et al., 2016). The spatiotemporal scales of the CZ processes range from the pore scale to the continental scale and from the geological past to the present and into the future (Brantley et al., 2007; Lin, 2014).

Such interdisciplinary and multiscale study of terrestrial ecosystem processes may be best accomplished through co-located Critical Zone Observatories (CZOs) where multiple scientific communities study various aspects of the CZ that can lead to synthesized understanding of complex systems (Anderson et al., 2008; Lin et al., 2011; Bernasconi, 2014; Brantley et al., 2016). It is the integration of Earth surface processes (such as landscape evolution, weathering, hydrology, geochemistry, and ecology) at multiple spatial and temporal scales and across anthropogenic gradients that is key to the concept of CZ science and CZOs. However, various conceptions of a CZO exist, and a more concretized concept of a CZO (e.g., how to separate a CZO from a long-term field site) is still an open issue in the scientific community.

In 2007, the first three national CZOs were established in the United States through funding from the US National Science Foundation. Through a decade of efforts, there are now 10 CZOs in the United States, spanning a wide range of climatic, ecologic, geologic, and anthropogenic environments (Fig. 1). Similar efforts were pursued in Europe, including four Terrestrial Environmental Observatories (TERENO) established in Germany in 2008 and the European Union’s four CZOs launched by Soil Transformation in European Catchments (SoilTrEC) in 2009 (Fig. 1). In China, a set of seven CZOs is being developed, among which five were co-funded by a Sino-UK joint program in 2015. In Australia, several CZOs are being established, especially in association with the Long-Term Ecological Research Network sites that are linked to the Terrestrial Ecosystem Research Network (TERN) created in 2009. Based on a summary from Banwart et al. (2013), a total of 69 CZO-like sites have been registered worldwide that can help forge independently conceived CZOs or CZO-like sites into a global network (Fig. 1).

Given the growing recognition of the significance of CZ research and observatories, the aim of this focus-topic update is to survey the latest CZ-related studies and highlight new accomplishments and emerging concepts in CZ science. We address the following two themes: (i) recent results and emerging concepts in CZ science under a conceptual framework of “deep” science, and (ii) knowledge gaps and major opportunities in advancing the frontiers of CZ science. We are hopeful that such an update can facilitate continuous discussions and stimulate further advancement in CZ science around the world.

The special section titled “Critical Zone Observatories” published in Vadose Zone Journal in 2011 was the first collection of studies on CZOs published in a scientific journal, marking the beginning of broader interests in CZ science. From 2011 to 2015, more than 200 CZ-related studies have been published in Vadose Zone Journal and other journals (Fig. 2) based on the ISI Web of Knowledge. These studies comprise interdisciplinary efforts from soil science, hydrology, biogeochemistry, geomorphology, geophysics, ecology, and other fields, which have advanced our understanding, prediction, and management of the CZ. In the following, we highlight the main findings from these CZ-related studies.

Critical Zone science would be too dispersive and complicated to understand if we outline recent CZ results by each discipline involved. Instead, we advocate a framework of “deep” science to help organize and comprehend research done in CZ science with a more synergistic perspective (Fig. 3). Three foci are included in this perceived “deep” science framework: deep time, deep depth, and deep coupling (Fig. 3). This “deep” science concept highlights the essence of integrating Earth surface processes at multiple spatial and temporal scales and signifies the unique contributions of CZ science to environmental and ecological research.

Processes within the CZ take place across temporal scales that ranges from multimillion-year time frames of tectonics to rapid transformation of short-term events like C flux and water cycling (Brantley et al., 2007). Recent CZ studies have suggested that the present landscape has been shaped through geologic time frames, and such geologic history (as recorded in the CZ) can provide scientific evidence and decision support to help project future CZ changes (Banwart et al., 2013).

The CZ processes may be grouped into (i) fast cyclic processes (e.g., diurnal fluctuation of soil temperature, seasonal changes in soil moisture, and yearly changes in vegetation growth) and (ii) long-term cumulative changes (such as bedrock weathering, pedogenesis, and ecosystem succession). The shorter and longer time scale CZ processes are intertwined, with interactions, feedbacks, thresholds, and cumulative effects (Lin, 2011). For example, it takes hundreds of thousands or more years for bedrock to be transformed into soil. These slow cumulative weathering and pedogenic processes lead to specific soil structures that control water movement in the soil profile and the periodic changes in soil moisture. On the other hand, each pulse of water moving through the soil profile causes varying degrees of physical translocation of materials, chemical reactions, and/or biological responses, thus imprinting certain marks of change in the soil profile. The cumulative effect of small changes over long periods of time can and do cause noticeable and permanent changes in soil evolution and soil structure (Lin, 2011).

The special section “Soil Architecture and Functions” in Vadose Zone Journal in 2012 emphasized the need to link soil architecture, a result of longer time scale processes, to shorter time scale soil physical, chemical, and biological functions (de Jonge et al., 2012). Studies in this special section explored major biophysical parameters and drivers for soil architecture and soil functions (e.g., Hamamoto et al., 2012; Subedi et al., 2012) and integrated emerging methods and techniques for assessing soil functions and visualizing soil architecture and its evolution in time and space (e.g., Markgraf et al., 2012; Sammartino et al., 2012).

Based on multidisciplinary studies of a 150-yr soil chronosequence at the Damma Glacier CZO in Switzerland, Bernasconi et al. (2011) found that the rapid evolution of microbial and plant communities strongly affected longer term weathering rate and soil formation. Biological variables indicated that the ecosystem in this glacier CZO evolved from sandy soils with a few plants to an ecosystem with almost complete vegetation cover in less than 70 yr, and to soils with a clear structure within 100 yr. The higher biomass and biodiversity in the older soils lead to more extensive biologic activities and faster pedogenesis. At the Koiliaris CZO in Greece, Moraetis et al. (2015) investigated the impacts from longer term natural processes and shorter term human activities on sediment provenance and soil formation. They found that longer term sediment transport from outcrops and strong eolian input from upslope, together with shorter term land use changes (e.g., man-made terraces), induced various soil development processes. At the Shale Hills CZO in Pennsylvania, Liu and Lin (2015) found that soil-terrain attributes formed over geological time scales played a key role in controlling the dynamics of fast subsurface preferential flow. Shi et al. (2015) suggested that detailed information on longer time landscape evolution and soil formation, such as soil type and soil thickness, were key to modeling day-to-day variations in soil moisture patterns at the Shale Hills CZO.

Evolution of the CZ entails a combination of threshold changes and gradual changes. Threshold pedogenic processes have been noted across contrasting soils and climates (Vitousek and Chadwick, 2013). Thresholds may be reached either through gradual changes when cumulative effects exceed a certain level (i.e., an internal or intrinsic threshold, such as element depletion) or due to rapid exogenous or endogenous change, when the rate is faster than the adaptive capacity of the system (i.e., an external or extrinsic threshold, such as human disturbance) (Lin, 2011). For example, based on a paddy soil chronosequence consisting of five profiles with cultivation history from 0 to 1000 yr, Chen et al. (2011) found that anthropogenic activities caused rapid changes in soil organic C (SOC), Ca, Na, and Mg contents within 50 yr, exhibiting external thresholds, whereas the changes in horizon differentiation of Fe oxide and clay illuviation during the 300- to 700-yr time period indicated internal thresholds for carbonate leaching.

The application of the threshold concept in CZ studies can propel our understanding of CZ processes and evolution. Some efforts have been pursued to investigate the causes of abrupt, rapid, and irreversible changes (i.e., thresholds) in CZ properties. At the Luquillo CZO in Puerto Rico, Porder et al. (2015) tested competing hypotheses (pedogenic thresholds, i.e., spatially sharp transitions in soil properties, vs. landscape homogeneity) on two parent materials (rock types). They found that, whereas the landscape underlain by volcaniclastics did not exhibit a regional knickpoint, strong knickpoints were observed in streams underlain by quartz diorite. Their results suggested that regional uplift, lithology, weathering, atmospheric inputs, and forest communities together controlled the spatial threshold distribution in soil cation availability (i.e., a significant difference above vs. below a knickpoint). At the Wüstebach CZO in Germany, Wiekenkamp et al. (2016) compared the frequency of occurrence of subsurface preferential flow across the catchment after different precipitation events. They identified the existence of a spatially variable threshold of precipitation on preferential flow initiation. Such activation thresholds for preferential flow response to precipitation were also observed by Dusek and Vogel (2014), who modeled subsurface lateral flow above a shallow semipermeable soil–bedrock interface. In a related modeling effort, Nimmo (2016) developed a process-based model to predict the occurrence of subsurface preferential flow in macropores by considering the spatial distribution of soil matrix infiltrability as a threshold. In this model, an elementary matrix area was used to represent the local infiltrability of the soil matrix material between macropores. After each elementary matrix area absorbed water up to its matrix infiltrability (i.e., the threshold value), excess water flowed into a macropore and initiated preferential flow.

Knowledge of CZ structure and its evolution in time and space is central to understanding CZ processes and predicting CZ response to future changes in climate and land use. Some efforts have been made to predict soil formation rate from bedrock porosity, matric potential, permeability, and mineralogy; to predict landform evolution from initial conditions and climatic, tectonic, and lithological forcing; and to predict CZ services from land use and management variations (Banwart et al., 2013). For example, Todorovic et al. (2014) predicted SOC stocks at the 0- to 20-cm soil depth at the Fuchsenbigl CZO in Austria by using an improved and calibrated RothC-26.3 model. The model results indicated that a change in land use from forest to grassland and cropping would result in a clear decrease in the amount and quality of SOC. At the Catalina-Jemez CZO in Arizona and New Mexico, Zapata-Rios et al. (2015) applied the effective energy and mass transfer (EEMT) concept to predict water transit times (WTT). Significant correlations were identified between EEMT and WTT, which suggested that basic climatic data embodied in EEMT could help predict hydrological and hydrochemical responses. Hunt (2015) established a scaling theory of solute transport on percolation clusters to predict the thickness of weathering rinds as a function of time. Given the general consistency of the relationship between weathering rind thickness and permeability, this study suggested the possibility of establishing specific models of weathering rind development according to rock type, grain size, permeability, and ambient conditions.

To support research at the Shale Hills CZO, a group of modules for simulating coupled processes in the CZ have been developed within the framework of the Penn State Integrated Hydrologic Model (PIHM) (Duffy et al., 2014). Flux-PIHM, Flux-PIHM-BGC, PIHM-SED, and Regolith-RT-PIHM have been or are being developed to model water and energy fluxes, C and N fluxes, sediment transport, and reactive transport, respectively, across a wide range of temporal scales from minutes to millions of years (Shi et al., 2014; Duffy et al., 2014). Given the drawback of current landscape evolution models that could not take into account realistic groundwater and overland flow and channel–hillslope interactions, Zhang et al. (2016) integrated hydrologic processes with hillslope and channel sediment transport processes to establish a new hydrologic-morphodynamic model, called the 3-D LE-PIHM. Their study (Zhang et al., 2016) tested the performance of LE-PIHM in modeling weathering and landscape evolution. Based on an ~2 × 10⁴–yr simulation, results from LE-PIHM indicated the importance of coupling groundwater flow in landscape evolution modeling. With the improvements in such an array of models, it should become possible to project the CZ under future scenarios of climate and land use.

The concept of the CZ extends the traditional perception of soils (i.e., 1–2 m deep with a focus on the root zone, as emphasized in ecosystem and agronomic studies) to often much deeper weathered bedrock. This vertical integration of Earth surface components provides a more holistic view of terrestrial physical, chemical, and biological processes and offers a long-term understanding of CZ services to human society. Efforts are underway to understand the linkages between the rhizosphere and the deeper subsurface (lithosphere) and between aboveground dynamics and belowground processes (e.g., Naithani et al., 2013; Hahm et al., 2014; Stone et al., 2014, 2015; Gaines et al., 2015, 2016). In the meantime, it has been recognized that upscaling from field plot measurements to hillslope, catchment, and watershed scales is crucial to understanding CZ processes. In general, recent CZ studies have suggested viewing the surface, shallow, and deep subsurface as a unified system to more fully understand the processes, structures, and functions of the CZ across different spatial scales.

As highlighted by Graham et al. (2010) and Brantley (2010), chemical, physical, and biological processes combine to transform fresh bedrock (granite) surrounded by only a few pores or fractures into saprock and eventually weathered rock and then soils surrounding small, isolated rock fragments. Linking parent material to soil properties, soil microbial communities, and vegetation distributions is an active research topic in CZ science. For example, at the Luquillo CZO, Stone et al. (2015) found differences in the dominant bacterial community structure with depth for two contrasting parent materials and two forest types within the upper 1.4 m of soil profiles. That study indicated that in tropical forests, bacterial communities exhibit the capacity to perform N-cycle transformations in deeper parts of the soil profile. At the Southern Sierra CZO in California, Hahm et al. (2014) found that differences in forest cover could be explained by the varying geochemical composition of the underlying bedrock. That study indicated that, in addition to climatic regulation, bottom-up lithologic control could play a key role in the distribution and diversity of vegetation in granitic mountain ranges. At the Eel CZO in northern California, Salve et al. (2012) found that the weathered bedrock zone created a hydrologically active domain to conduct substantial rainfall penetrating to depth and recharge a groundwater table perched on the underlying fresh bedrock. During the summer dry season, the shallow weathered bedrock zone dried much more slowly than the soil layer, which would influence water delivery to depth and water availability to vegetation. In addition to soil water storage, the weathered bedrock zone was an essentially unmapped moisture reservoir at the study site.

The application and advancement in geophysical techniques have enhanced our ability to image and investigate the deep subsurface and promote CZ science in linking deep weathering processes to shallow subsurface and surface processes. Parsekian et al. (2015) reviewed geophysical methods for examining CZ form and function across multiple spatial scales, from centimeters to kilometers. Their study identified three advantages of geophysical measurements as a necessary complement to direct observations obtained by drilling or field sampling: (i) noninvasive imaging of the geometry of structural features in the space between direct measurements; (ii) repeated data acquisition in real time and with time in the field, and (iii) relatively larger spatial coverage, beyond what direct measurements can reveal, that can be compared with data sets from remote sensing techniques.

We have selected several studies to illustrate the strength of geophysical methods in advancing CZ science. St. Clair et al. (2015) applied seismic velocity and electrical resistivity surveys in three landscapes within the US CZO network (Boulder Creek, Calhoun, and Christina CZOs) and found that the pattern of bedrock fracture distribution as it neared the surface could not be easily explained by climate but probably depended on topographic stresses. Their study demonstrated the utility of geophysical methods to challenge traditional understanding of bedrock disaggregation, groundwater flow, chemical weathering, and the depth of the CZ (Anderson, 2015). At the Southern Sierra CZO, Holbrook et al. (2014) used seismic refraction and resistivity imaging to detect variations in soil thickness and porosity across the catchment. Their results suggested that saprolite is a crucial reservoir for water storage, where the residence times of soil in the Southern Sierra are on the order of 10⁵ years, indicating possible integrations of weathering over glacial–interglacial fluctuations. At the Shale Hills CZO, Guo et al. (2014) combined time-lapse ground-penetrating radar (GPR) surveys and artificial infiltration to map the pathways of subsurface lateral flow for the upper 1.4 m of soil profiles. They reported that enhanced radar data post-processing procedures provided a means to reconstruct the subsurface flow network and its dynamics with time. Zhang et al. (2014) compared the seasonal changes in GPR signals in two contrasting soils at the Shale Hills CZO. They found that repeated GPR surveys could identify the impacts of soil layering on water flow and water distribution in soils.

Increasing efforts have arisen in the CZ science community to link aboveground and belowground systems and to improve our understanding of how the surface and the deep subsurface are connected. These connections include: (i) water flux through the soil–plant–atmosphere continuum (SPAC); (ii) the influence of precipitation partitioning by canopy interception on subsurface preferential flow occurrence; (iii) interaction between vegetation distribution and soil moisture spatial pattern; and (iv) surface and subsurface flow networks at the hillslope and catchment scales (Li et al., 2012).

The special section titled “Soil–Plant–Atmosphere Continuum” in Vadose Zone Journal in 2012 reported advances in the measurement and understanding of soil–water, soil–plant, and plant–atmosphere interfacial processes. A series of interdisciplinary studies were included in this special section that linked aboveground and belowground processes, including the new methods of quantifying and modeling soil–plant interactions from a single plant to the field scale (e.g., Zarebanadkouki et al., 2012; Assouline et al., 2012), the impact of canopy interception and root uptake on the variability of soil moisture in space and time (e.g., Moradi et al., 2012; Guswa, 2012), modeling water flux through the SPAC (e.g., Romano et al., 2012; Schröder et al., 2012), and the improved assessment of soil moisture and land–atmosphere interactions (e.g., Scanlon and Kustas, 2012). Although the synthesis of knowledge from various compartments of the SPAC into a holistic view remains challenging, these studies brought a new perspective that the whole CZ could function as a complex system.

The Shale Hills CZO, covered by temperate forest and underbrush, provides an ideal experimental site to study the interaction between aboveground ecohydrological and belowground hydropedological processes. Naithani et al. (2013) combined data sets of soil water content measurements (using time-domain reflectometry [TDR]), ground-based and satellite-sensed leaf area index, the spatial distribution of the dominant tree species, and high-resolution 0.5- by 0.5-m digital elevation model to evaluate the impacts of canopy growth on soil water content. They found that the spatial patterns of vegetation and soil moisture become increasingly homogenized and coupled from leaf onset to maturity but heterogeneous and uncoupled from leaf maturity to senescence. Based on multiple lines of evidence, including stable isotope natural abundance, sap flux, and soil moisture depletion patterns with depth, Gaines et al. (2015) inferred that root water uptake mainly occurred in the shallow (less than ~60 cm) soils throughout the Shale Hills CZO, even during the dry period of the growing season. Gaines et al. (2016) used a deuterium tracer, sap flux, and anatomical techniques to study tree water transport on a ridgetop at the Shale Hills CZO. They found that the soil-to-leaf driving force for water transport may control tracer velocity across tree species. Kraepiel et al. (2015) and Meek et al. (2016) conducted a multi-tracer study at the Shale Hills CZO to study the source and cycling of nutrients and metals among atmosphere, bedrock, soil, stream, groundwater, and vegetation. Their results indicated a strong interaction of slower weathering processes with rapid, biologically driven cycling between soils and vegetation.

In addition to the vertical integration of CZ processes from the rhizosphere down to the weathered bedrock and up to the aboveground canopy, the concept of deep depth also views the whole CZ as a unified system and extends field measurements to a larger spatial scale, such as catchment and watershed scales. The special section titled “HOBE: A Hydrological Observatory” in Vadose Zone Journal in 2011 reported findings from the Skjern catchment in Denmark and recognized the need to understand CZ processes at the catchment scale (Jensen and Illangasekare, 2011). Studies in this special section reported on hydrological processes at different spatial scales, including regional estimates of precipitation from Doppler radar and its impact on the catchment-scale modeling of hydrological processes (e.g., He et al., 2011; Fu et al., 2011), measurement of evapotranspiration (ET) and greenhouse gas emissions and their dynamics and underlying controls at the catchment scale (e.g., Ringgaard et al., 2011; Herbst et al., 2011), and the impacts of climatic changes on hydrological processes at the regional scale (e.g., van Roosmalen et al., 2011).

High-density instrument arrays installed in CZOs can provide time series measurements that cover a whole catchment and help study CZ dynamics in a more comprehensive manner. For example, recent advances in sensor networks have created new opportunities to study soil water, soil temperature, soil matric potential, and other CZ variables with high temporal resolution and at the catchment scale. In the following, we use soil water content and soil matric potential as examples to illustrate the efforts of the CZ scientific community working toward a holistic view of CZ processes at the catchment and larger spatial scales.

Soil water content is a key variable in many processes within the CZ. The characterization of spatial soil water patterns and their temporal evolution is a cornerstone of our understanding of hydrological fluxes, flow pathways, and their impacts on biogeochemical dynamics. Using a 14-mo data set collected from 240 sensors across a 1.44-km² catchment at the Schäfertal CZO in Germany, Martini et al. (2015) characterized the spatial pattern of soil water content and inferred the hydrological processes that probably controlled the formation and dynamics of such a pattern. The results of that study indicated that the spatial variability of soil water content decreased with soil depth and that a variety of hydrological processes occurred at different times and at different topographic positions and soil horizons, thus the topsoil’s wetness might not necessarily reflect the hydrological processes occurring within the soil profile. This study pointed out the difficulty of upscaling soil water content from point measurement to the catchment scale. At the Rollesbroich CZO in Germany, Qu et al. (2014) investigated the spatiotemporal variability of soil water content patterns using a 2-yr time series measured at 41 locations across the 0.3-km² catchment. Temporally stable characteristics were found in the spatial variability of soil water content, which was correlated with the spatial variation in hydraulic properties. Webb et al. (2015) used a 3-yr data set collected from 97 sensors across 27 locations at the Southern Sierra CZO to investigate the wetting and drying dynamics of the shallow subsurface during snowmelt. Their study found a high variability of wetting and drying dynamics within the top 1 m of the soil in this mixed-conifer forest at the sub-hillslope and catchment scales, which could be influenced by subsurface lateral flows developed in the shallow soil or along the soil–bedrock interface after snowmelt. Given the cost of developing and maintaining a soil moisture sensor network, Schröter et al. (2015) used a terrain-based fuzzy c-means sampling and estimation approach at the Schäfertal CZO to identify representative measurement locations by stratifying the landscape into unique combinations of topographic attribute values. Their study demonstrated the effectiveness of the point-scale measurements in revealing the spatiotemporal pattern of soil moisture in a relatively small catchment.

While the spatiotemporal characteristics of soil moisture content have been extensively investigated, a limited number of studies have attempted to understand the spatiotemporal organization of the soil matric potential, especially at the catchment scale. Based on a 5.5-yr data set collected from 62 sites by standard nested tensiometers across the Shale Hills CZO, Yu et al. (2015) investigated the spatial variability and temporal stability of the soil matric potential at multiple depths (10–100 cm). They found that the spatiotemporal structure of the soil matric potential in the catchment was controlled by soil type, topography, season, and soil depth. Their results indicated that the downward parabolic trend in the spatial variability of the soil matric potential occurred when the spatial mean value of the soil matric potential decreased (i.e., became drier) across all soil depths.

In addition to the catchment or watershed scale, some studies have examined the spatiotemporal pattern of soil water content at the regional and even national scale. For example, based on the soil water content data obtained from the Soil Climate Analysis Network (SCAN), Wang and Franz (2015) assessed different climatic and soil controls on the soil water content spatial variability in Utah and the US Southeast. Their results suggested that instead of being controlled by climatic variables at the regional scale (~10⁵ km²), as traditionally believed, the spatial variability of soil water content in the study areas was strongly dependent on soil hydraulic properties. This regional-scale study also provided important implications for verifying remotely sensed soil moisture data (e.g., the Soil Moisture Active Passive mission [SMAP]) and initializing and parameterizing regional land surface and climate models. During 2009 to 2011, the US Climate Reference Network (USCRN) deployed Hydraprobe sensors to monitor soil water content across the United States. Bell et al. (2015) used the soil water content measurements from 114 USCRN sites at all observation depths to evaluate the 2012 drought. They detected an overall 11.07% decrease in soil water content from the average of the 2011 to 2013 summers. Their study demonstrated the utility of the USCRN for monitoring national soil water conditions, evaluating droughts, and tracking climate change with time. Coopersmith et al. (2016) quantified the random error of soil water content measurements from the USCRN. Based on the data from all of the ground sites (each site with three Hydraprobe sensors collocated), their results indicated that the average random error of Hydraprobe measurements was 0.012 m³/m³. The random error magnitudes of three repeated sensors at each ground site were related to precipitation patterns.

The integrated concept of the CZ reflects a recognition that the interactions and feedbacks among geologic, pedologic, hydrologic, chemical, biologic, atmospheric, and anthropogenic processes are coupled, especially those mediated by the flux of freshwater. Such close interplays among CZ processes require a synergistic approach from multidisciplinary and interdisciplinary perspectives to fully understand the co-evolution of CZ structures and functions.

A wide range of physical, chemical, and biological processes interact within the CZ. This diversity of interactions is central to understanding CZ evolution and function. Among various couplings, biogeochemical and hydropedological processes are ubiquitous across CZOs. The special section titled “Critical Zone Observatories” in Vadose Zone Journal in 2011 reported the initial results from the first several CZOs, which attempted to couple hydrology, biogeochemistry, and their interplay in the CZ (Lin et al., 2011). For example, at the Boulder Creek CZO in Colorado, Dethier and Bove (2011) suggested that clay and Fe-oxide mineralogy could be used to distinguish alterations caused by weathering from that produced by hydrothermal activity. Chorover et al. (2011) developed a conceptual framework for quantifying the EEMT and found that the coupled C and water fluxes controlled the evolution of the CZ at the Catalina-Jemez CZO. At the Shale Hills CZO, Jin et al. (2011) showed that Mg concentration dynamics were controlled by the kinetics of clay mineral dissolution and soil exchange capacity, whereas changes in water isotopologues were strongly related to water residence times. They also found that rock weathering mechanisms and hydrochemical measurements could be used to infer hydrologic processes. Also at the Shale Hills CZO, Andrews et al. (2011) compared SOC storage and soil pore water dissolved organic C (DOC) across various landform units and reported that the higher SOC storage and soil pore water DOC concentration in swales were probably related to active subsurface hydrological processes in the swales because they were transport-driven hot spots of organic C in the studied catchment. These examples illustrate the importance of coupled biogeochemical and hydropedological processes in CZOs.

The special section titled “Frontiers of Hydropedology in Vadose Zone Research” in Vadose Zone Journal in 2013 investigated the two-way interactions between soil architecture and water movement (Vogel et al., 2013). This special section highlighted hydropedology as a new scientific approach that emphasizes the central role of water in a variety of soil processes, including solute transport and various biogeochemical processes (Vogel et al., 2013). For example, the spatiotemporal investigation of high-resolution sampling of stable isotopes (δO and δD) suggested the prevalence and mechanisms of preferential flow at the Shale Hills CZO (Thomas et al., 2013). That study suggested the potential of using soil water isotopes to connect biogeochemical and hydropedological processes. Hardie et al. (2013) examined the dependence of subsurface flow patterns on the soil structure for soils of contrasting textures in Tasmania, Australia, and discussed the possible impacts of preferential flow on the eluviation of clay and chemical reduction in shallow soils. Arnold et al. (2013) examined fundamental hydropedological and ecohydrological relationships in natural Brigalow ecosystems of eastern Australia to support the rehabilitation of disturbed semiarid environments by promoting the development of native plants.

Dissolved organic matter (DOM) plays an important role in many biogeochemical and hydropedological processes that take place in the CZ. Variations in the quality or quantity of stream-water DOM can be used as a proxy of changes in land use or as guidance to catchment management. The special section titled “Dissolved Organic Matter in Soil” in Vadose Zone Journal in 2014 highlighted the role of DOM in C and nutrient cycling, pedogenesis, and microbial metabolism that link the CZ to aquatic systems (Jansen et al., 2014). Studies in this special section improved the conceptual understanding of DOM dynamics in the CZ and aquatic systems by tracing changes in DOM concentration along its transport path from the soil to surface water (e.g., Roth et al., 2014; Vázquez-Ortega et al., 2014), explored the interactions of DOM with hydrology to build the linkage between the CZ and aquatic systems (e.g., Klotzbücher et al., 2014; Van Gaelen et al., 2014), and assessed the interrelations between DOM dynamics and human impacts and their possible changes in the future (e.g., De Troyer et al., 2014; Thangarajan et al., 2014). In addition to this special section, many DOM-related studies have been conducted across different CZOs. For example, Bol et al. (2015) used a 4-yr data set of weekly DOM features, pH, and Fe content sampled in the streams of the Wüstebach CZO to investigate the spatiotemporal variation of DOM. The positive correlation between DOC and Fe concentrations in the water sources studied indicated the transport of DOM via organo-mineral complexes. That study demonstrated the value of a long-term spatiotemporal data set covering DOM quality and quantity in both tributaries and main-stream water, as they apportion contributing sources and drivers of DOM dynamics in sub- and whole-catchment stream water. All these studies suggested the potential of DOM in coupling biogeochemical and hydropedological processes and promoting the frontier of CZ science.

Due to the closely intertwined CZ processes occurring across a wide range of spatial and temporal scales, multidisciplinary and interdisciplinary techniques and expertise are required for comprehensive understanding of CZ dynamics. Remote sensing techniques have provided many new ways to monitor and estimate various CZ attributes across multiple scales, which can enhance CZ science. The special section titled “Remote Sensing for Vadose Zone Hydrology” in Vadose Zone Journal in 2013 demonstrated the improved temporal and spatial quantification of soil water content, ET, soil hydraulic parameters, soil salinity, and vegetation dynamics. The techniques use optical, microwave, gravitational, infrared, and other sensors (Mohanty et al., 2013). A total of 14 applications of remote sensing techniques to study the CZ were included in this special section, aiming to (i) improve the estimation, scaling, and data assimilation of soil water content and its variability by microwave remote sensing (e.g., Han et al., 2013; Chaouch et al., 2013), (ii) enhance the estimation of ET and vadose zone properties by multispectral satellite data (e.g., Landsat, Moderate Resolution Imaging Spectroradiometer [MODIS] and Advanced Very High Resolution Radiometer [AVHRR]) and microwave remote sensing (e.g., Gowda et al., 2013; Shin et al., 2013; Teluguntla et al., 2013), and (iii) ground truth soil water content to calibrate and validate microwave remote sensing retrievals (e.g., Dorigo et al., 2013). Findings in this special section suggested the potential of remote sensing techniques in CZ research, including the investigation of land–atmosphere interactions, hydrology, water resource management, and hazard assessment (Mohanty et al., 2013).

Besides the studies included in this special section, Harpold et al. (2015) reviewed the application of lidar technology in obtaining precise three-dimensional information on the Earth’s surface characteristics and described the unique features of lidar data sets in advancing CZ research. For example, terrestrial laser scanning was used to reveal the important influence of alluvial and colluvial processes on vegetation and soil dynamics on semiarid hillslopes (Harman et al., 2014). Without the high resolution and precision of lidar information on surface topography and canopy volume, it is very difficult to quantitatively investigate the co-evolution of microtopography and vegetation (Harman et al., 2014). Lidar-derived canopy structure was used to test a CZ development model based on eco-pedo-geomorphic feedbacks and to identify the crucial role that ecological processes (e.g., vegetation density) have played in landscape evolution (Pelletier and Perron, 2012). These exemplar CZ studies elucidated the potential of lidar to simultaneously quantify topographic, vegetative, hydrological, and other CZ processes (Harpold et al., 2015). Harpold et al. (2015) also identified three areas where lidar technology can further the frontier of CZ science: (i) quantifying landscape adjustments through time, (ii) parameterizing and verifying CZ models, and (iii) improving the understanding of coupled CZ processes across multiple scales.

In addition to using single technologies to monitor CZ processes across multiple scales (e.g., lidar), increasing efforts are being made to integrate different techniques to study CZ dynamics. For example, at the Koiliaris CZO, Tsiknia et al. (2014) applied molecular biotechnology, soil physical, chemical, and biochemical analyses, and geostatistical mapping to investigate the underlying factors that shape the spatial distribution of dominant soil microorganisms in the watershed. At the Luquillo CZO, Stone et al. (2014) integrated soil chemical analyses, phospholipid fatty acid analyses, and enzyme assays to determine how extracellular enzyme activity changes as a function of soil depth. Moreover, some studies have compared the consistency of different technologies in measuring the same CZ variable. For example, at the Reynolds CZO in Idaho, Flerchinger and Seyfried (2014) compared several methods for quantifying ET during an 8-yr period, including eddy covariance systems with and without adjusting the turbulent fluxes to force energy balance closure, soil water storage loss measured by TDR and neutron probes, and simulation by the Simultaneous Heat and Water (SHAW) model. Their results suggested that, compared with other approaches, ET estimates from eddy covariance significantly underestimated seasonal ET unless the turbulent fluxes were adjusted to force the closure of the energy balance. In a 0.1-km² forest catchment, Lv et al. (2014) compared soil water content measured using a cosmic-ray neutron probe (CRNP) and an in situ time-domain transmissometry (TDT) sensor network from a total of 108 sensors. Additionally, the near-surface soil water content was simulated using a HYDRUS-1D numerical model. Then the simulated and TDT-measured soil water content were used to construct the depth-weighted mean areal soil water content for comparison with the CRNP measurements. During a period of 6 mo, the CRNP estimates exhibited a dry bias under relatively wet conditions at the beginning of the snow-free period. Using a combination of soil water content measurements and near-surface simulations, the CRNP output was recalibrated to capture the wetter conditions. This study indicated possible discrepant soil moisture content from different devices and calibration methods, which calls for attention to the needed standardization in methods, equipment, protocols, databases, and models to maximize the potential of a global network of CZOs. The studies mentioned above indicated the importance of advancing and integrating multidisciplinary and interdisciplinary techniques in achieving a more comprehensive understanding of the CZ. In addition to the combination of established observation methods, the interdisciplinary aspect also requires developing new concepts and methods that can study the CZ as a whole system to integrate results from multidisciplinary observation methods and to provide systematic understanding of CZ processes.

Although widespread recognition that CZ processes have important societal relevance, depletion of natural resources from the CZ is accelerating, and in many cases the CZ is managed without regard for the limits of supporting ecosystems. The intensified pressure on the CZ calls for sustainable management of the CZ (Robinson et al., 2013). To facilitate the application of CZ science to sustainability practices, a new concept of CZ services was proposed by Banwart et al. (2012) and Field et al. (2015). Banwart et al. (2012) conceptualized mass and energy flows that arise from CZ processes that supply goods and services for human benefit. Field et al. (2015) presented a CZ perspective to complement and enrich the current perspectives on ecosystem services in terms of context, constraints, and currency. Understanding ecosystem services within the context of CZ services requires enhanced efforts to address a grand challenge in integrating bio- and geosciences. The concept of CZ services integrates biological services with soil generation, landscape evolution, and water cycling, emphasizing geological processes and nonrenewable resources beyond the scale of a human life span, which extends the biologically focused perspectives of ecosystem services (Field et al., 2015).

In essence, CZ services extend the context for ecosystem services in three important aspects: (i) explicitly addressing how the physical structure of the terrestrial surface (e.g., parent material, topography, and orography) provides a broader spatial and temporal template that determines the co-evolution of physical, chemical, and biological processes in ecosystems; (ii) emphasizing the rate-limiting processes of ecosystem services that are fundamentally constrained by CZ processes, such as soil formation, nutrient supply, hydrologic partitioning, and streamflow generation; and (iii) integrating CZ processes into an evaluation currency by quantifying the amount of EEMT (i.e., a measure of the energy flux available to do physical, chemical, and biological work within the CZ) (Field et al., 2015). This proposed concept of CZ services builds a linkage between ecosystem services and the constraints thereon associated with CZ processes and thus provides a framework to enable more effective long-term management and valuation of ecosystem services in response to a changing climate and increasing human disturbances (Banwart et al., 2012; Field et al., 2015).

Given the vital influence of the CZ to life and humanity, CZ management can no longer be based on a single function but instead should be coupled with multiple services that the CZ offers (Robinson et al., 2013; Baveye et al., 2016). In particular, with the recognition that mankind has entered the new Anthropocene era in which human activities substantially impact Planet Earth, research outcomes from CZ science should provide decision support to evaluate different adaptive strategies for mitigating the impacts of climatic change and human disturbance so as to sustain CZ services (Banwart et al., 2013; Montanarella and Panagos, 2015; Baveye et al., 2016). Montanarella and Panagos (2015) demonstrated how the CZ can be used as a new paradigm to contribute to environmental policy decisions, including climate change, water management, biodiversity protection, air quality, water quality, waste management, and agricultural policies. This study suggested that when policymakers make decisions on land use, water management, and agricultural practices, they should consider soil threats and soil functions and the trade-off between increasing food production vs. organic C loss or pollutant transformation. For example, at the Intensively Managed Landscapes CZO inn Illinois, Iowa, and Minnesota, Papanicolaou et al. (2015) compared the spatial variability of soil saturated hydraulic conductivity (a key soil property for CZ service that influences the amount of runoff, erosion, and stream-water quality) at three sites with different agricultural management practices. They found that different land use and management practices (i.e., conventional tillage, no-till, and Conservation Reserve Program) triggered statistically significant differences in the soil saturated hydraulic conductivity. Their results can provide decision support to policymakers to consider the balance between agricultural activity and CZ services. Future efforts are required to translate new CZ science knowledge into policy-relevant data and information to support policy and management decisions related to CZ sustainability.

The past few years have seen a global network of CZOs beginning to take shape, offering a potentially fertile ground to extend the breadth and depth of interdisciplinary CZ science (Fig. 1). Up to this point, however, most CZ studies have focused on the outcomes from a single CZO, with a limited number of studies comparing several different CZOs. No study has been done to identify common catchment behaviors across CZOs, albeit that each CZO has its unique features. Given the growing interest in CZOs worldwide, a real challenge and opportunity for the CZ scientific community is to synergize outcomes from multiple CZOs and to synthesize data or knowledge via the network of CZOs to reveal the underlying controls on CZ structure, evolution, processes, and services.

This synergistic effort calls for establishing CZO standards to maximize the potential of a global array of CZOs. As the results reported by Flerchinger and Seyfried (2014) and Lv et al. (2014) showed, different measuring devices and calibration methods would probably lead to discrepant values of the same variables measured, which will hamper comparability across CZOs. Thus, more efforts are needed to standardize the protocols of CZO investigations in terms of observation variables and processes, measurement equipment and methods, database organization and analysis, and model development and intercomparison. In addition to unique features and site-specific processes in individual CZOs, common infrastructure and baseline measurements of a CZO may include: (i) real-time monitoring of energy, water, solutes, and sediment fluxes across CZ components and boundaries via eddy covariance systems, sensor networks, and other standardized observation equipment; (ii) isotopes and other tracers of water, particles, and chemicals to couple hydrological and biogeochemical processes; (iii) geophysical investigations of the three-dimensional architecture of the CZ and the characterization of CZ storages and changes across different temporal scales; (iv) documentation of the presence and role of the biota community in the CZ, especially the microbial community in the deep subsurface; and (v) open database access with adequate metadata and a standardized format to facilitate interdisciplinary data exchange and meta-analysis across CZOs (e.g., Niu et al., 2014). Development of cross-CZO collaborations, in particular, will contribute to such needed standards to meet the full potential of a global CZO network through joint efforts in various areas, such as sample collection and analysis, monitoring setup and instrumentation, and model development and comparison, among others.

Efforts thus far in developing a global network of CZOs have been mainly pursued in the United States, Europe (especially Germany, the United Kingdom, and France), China, and Australia. However, CZOs in other countries are relatively rare, in particular in Africa and South America (Fig. 1). Given the goal of the global network of CZOs to develop predictive knowledge across terrestrial ecosystems, more efforts are required to reach out to other parts of the world to extend the spatial coverage of the global network of CZOs, on which comparative studies could be conducted across CZOs along environmental gradients or under different anthropogenic impacts to advance our understanding of the evolution, structure, and functions of the CZ.

Recent advances in Earth science have greatly enhanced our ability to develop Earth system models to represent a wide variety of processes that are coupled under different climatic conditions. Critical Zone science also creates new modeling opportunities to integrate Earth surface processes. Hence, similar efforts are needed to develop comprehensive CZ models or a library of models and databases (Vereecken et al., 2016). Two modeling needs have been recognized that may lead to substantial advancements in CZ science:

1. Models need to be developed and tested across CZOs with a focus on common CZ features. The recent findings in CZ science provide rich knowledge to develop a unified theoretical framework to understand CZ evolution and function, on which common models could be built to quantify basic CZ processes. These common models need to be tested across a variety of site conditions and should be more transferrable to conditions outside of those in which they were developed or in which extensive monitoring data are available and thus more capable of solving challenges associated with heterogeneity and regional differences. Extrapolating the common models to unmonitored sites and to access global CZ processes and services is essential to advance CZ science.

2. A common library of CZ models and databases should be established. Instead of one supermodel for all CZ processes, it may be more practical to create a library of models to include links between pedogenesis and landscape evolution, connections between aboveground and belowground, coupling of geophysics and geochemistry, integration of hydrology and biogeochemistry, and combination of anthropogenic and natural processes. A library of comprehensive CZ databases is also crucial, and better tools are needed for querying and sharing data sets with adequate metadata. These databases can then be linked to models and used for model development, calibration, and testing, as well as for hypothesis-driven investigations. To promote efficient data sharing and integration, more efforts are required to develop interdisciplinary data exchange platforms and to enhance the inter-operability of CZ databases across CZOs.

The ongoing funding-source-mandated open-data/open-model paradigm shift that is underway has improved the availability and accessibility of CZ data. The US CZO network has applied the Theory–Model–Data fusion framework to integrate CZ data across CZOs (http://criticalzone.org/national/data/). Currently, there are 287 CZ data sets publically available (as of July 2016). Online search and visualization tools have also been developed to help users to access CZ data. The Hydrologic Information System (HIS; http://his.cuahsi.org/) and HydroShare (https://www.hydroshare.org/) of the US National Science Foundation supported Consortium of Universities for the Advancement of Hydrologic Sciences (CUAHSI) are other examples of sharing data, models, and codes. On these two platforms, data publishers can organize, store, and make data available to others, and data users can search, retrieve, visualize, analyze, and run models on data from these data servers. Ongoing efforts by the CUASHI will further improve these platforms to become a standard system for storing, managing, organizing, indexing, documenting, and sharing data. Although online data sharing has great potential to improve CZ data accessibility, high-level integration of cyberinfrastructure is still needed to promote the information exchange among different data sharing systems. Because individual researchers can publish data on online platforms, data homogeneity (both data quality and data format) becomes an important issue for data sharing, which again calls for establishing data standards. Additionally, more efforts are needed to encourage individuals to share data through online platforms. Contributions to data publishing should be more recognized, such as assigning a DOI number to online published data sets, models, and codes, making them more widely citable and searchable.

Processes taking place within the Earth’s CZ are essential for sustaining life and humanity. Critical Zone science has been recognized as one of the most compelling research fields in the 21st century. The past few years have seen an increasing number of CZOs around the world that are being connected into a global network. This focus-topic update summarizes recently published work to highlight major results and new concepts related to CZ science, especially those generated from recognized CZOs. A framework of “deep” science—deep time, deep depth, and deep coupling—is used to synthesize recent CZ investigations across a broad range of spatial and temporal scales.

Deep time emphasizes the recognition of the intertwined short-term and long-term CZ processes and the linkage from past CZ evolution to current CZ functions to future CZ scenarios.

Deep depth provides a holistic view of the entire three-dimensional CZ as a unified system and suggests the need to integrate the surface and deep subsurface and to upscale field-plot measurements to hillslope-, catchment-, and watershed-scale understanding.

Deep coupling indicates the complex interplay of CZ processes (especially across biotic and abiotic processes) and suggests the need to link CZ services to CZ management.

The knowledge gaps and major opportunities suggested to advance the frontiers of CZ science are (i) a global network of CZOs and cross-CZO collaborations, and (ii) a library of CZ models and databases. This update should facilitate continuous discussions and stimulate further advancement of CZ science around the world.

The authors are supported in part by the US National Science Foundation Hydrologic Sciences Program (Grant EAR-1416881, PI: H. Lin) and the Critical Zone Observatory Program (EAR-0725019, PI: C. Duffy; and EAR-1239285, EAR-1331726, PI: S. Brantley). We thank Vadose Zone Journal Editor Harry Vereecken for this opportunity to update the community on the latest studies in Critical Zone research and observatories. We thank Jirka Šimůnek, Dave Stonestrom, Harry Vereecken, and another anonymous reviewer for their comments that have helped improve the quality of this paper.