Potential nitrogen mobilisation from the Yedoma permafrost domain

Since plant growth in high-latitude ecosystems is usually N limited (Beermann et al 2015, 2017), N release from thawing permafrost deposits may boost plant productivity and thus CO₂ fixation from the atmosphere through photosynthesis. However, the strength of this feedback is strongly affected by how well the liberated nutrients including N will be accessible for plants (Pedersen et al 2020) or microbes (Baker et al 2023). The accessibility of nutrients is influenced by the spatial and temporal patterns of N release and transport, the presence or absence of plants, plant, species composition, plant rooting depth as well as soil characteristics. Seasonal variability plays an important role here: Plants and microbes are able to immobilise available N (figure 2) before photosynthetic activity sets in during the winter/spring transition (Edwards and Jefferies 2010). However, seasonal declines in microbial biomass may result in another pulse of available N (Bardgett et al 2007). Active layer deepening, apart from exposing more OM to microbial decomposition, leads to a longer thawing period, thereby prolonging the period with biologically available soil N.

Active layer deepening likely also results in direct N inputs to the ecosystem N cycle from newly thawed permafrost to the bottom of the active layer, as well as longer soil N availability, especially in the first years after thaw (Salmon et al 2016, 2018).

It has been shown, however, that due to a temporal mismatch between peak vegetation growth (mid-summer) and maximum thawing depth (late summer/autumn), there is incomplete utilisation of nutrients by plants near the permafrost table (Lacroix et al 2022). Also, plants will not necessarily be able to access the N released from permafrost, if they are located deeper down than their rooting zone (Keuper et al 2017). Incomplete plant N use can cause a nutrient oversupply near the permafrost table, which may lead to freely available N for microbial cycling processes such as nitrification and denitrification. This, in turn, could lead to emissions of N₂O, a strong greenhouse gas and stratospheric ozone-depleting substance (figure 2 and Lacroix et al (2022)), or increased losses of N via downward vertical leaching and lateral flow (figure 2, chapter 3.2). This may be the case with gradual active layer deepening, while abrupt thaw exposes the newly thawed material to the surface, making it readily available for newly established vegetation on retrogressive thaw slumps (Marushchak et al 2021) or in thermokarst lakes (Walter Anthony et al 2014).

In retrogressive thaw slumps and other disturbed sites, however, development of full vegetation cover takes several years, and easily exceeds the capacity of plants to take up N during early stages of succession the mineral N supply (Marushchak et al 2021). This successional mismatch is probably also true for the process of surface water generation resulting from the melting of ground ice and subsequent subsidence, although there is limited literature on this topic. Thermokarst lake and gully formation leads to distinct vegetation dynamics, influencing the utilisation of newly available N by plants. The development of aquatic landscape features, such as small pools and troughs, requires colonisation by aquatic vegetation (primarily consisting of mosses and sedges) before they can effectively utilise the accessible N for carbon sequestration. Nevertheless, once the full vegetation cover is established on these disturbance sites, they exhibit the potential for considerable productivity.

Thus, increased N availability by permafrost thaw in soils could enhance net primary productivity and increase OC stocks in plant biomass, but this might not always be the case. Some experimental evidence confirms that plant growth increases with newly thawed deep N sources, while other studies show that effects on OC acquisition might be limited since only few deep-rooting plant species will be able to exploit the newly thawed N (Keuper et al 2017). Considering that a large part of the Yedoma domain lies within the continuous permafrost zone (with a shallow active layer depth of <1 m), it is likely that N in the soil layers exposed by active layer deepening may become accessible to deep rooting plants (see the gradual thaw pathway in figure 3), if there is no temporal mismatch.

Recently, a study of vegetation types in retrogressive thaw slumps (Buckeridge et al 2023) revealed that shrubs are the best candidate for plant N uptake as they have a deep vertical rooting potential (figure 2) and can reach the deep N stocks, as also found by Hewitt et al (2019). This is in line with the shrubification which has been observed in Arctic tundra due to recent warming (Frost et al 2013). Whether increased vegetation stocks may offset soil OC losses is therefore still uncertain (McGuire et al 2018, Mauritz et al 2021, Schuur et al 2021, 2022).

In the Yedoma domain (and other ice-wedge-bearing permafrost), the thawing of ground ice frequently results in elevated surface water levels. This phenomenon becomes apparent through the degradation of ice-wedge polygonal landscapes, resulting in the formation of polygonal pools and ice wedge troughs, and sometimes leading to interconnected hydrological pathways (Liljedahl et al 2016). Importantly, this process is likely to serve as an increasingly significant source of nutrients (Vonk et al 2015). For instance, Koch et al (2018) observed a rapid decline in ammonium (NH₄ ⁺) levels within deep trough ponds, attributable to biogeochemical cycling.

Runoff from Holocene cover on top of Yedoma uplands and some refrozen thermokarst basins sediments of the Yedoma domain can transport nutrients like NO₃ ⁻ and NH₄ ⁺ to aquatic systems. This runoff/lateral flow is limited in deeper strata, especially in Taberites beneath thermokarst lakes, as they are (at least partially) (re-)frozen. Vertical flow could occur as well if a groundwater connection is present, or through transport within plant roots in a through-going talik. Particulate N and DON originating from runoff could be deposited on floodplains, in river deltas, or in the nearshore sediments of the Arctic shelf seas (Ramage et al 2018, Marzadri et al 2021). If leached and transported to the aquatic systems, the fate of N strongly depends on the form and quality of N, as well as other stream network characteristics such as redox potential, oxygen status, availability of other nutrients, pH, temperature and microbial and plant community composition. Additionally, water transparency and sediment content are controlling the availability of light and thus nutrient uptake and primary productivity.

Nutrient losses through lateral flow and vertical leaching are potentially a major factor in overall tundra nutrient budgets (Mack et al 2004, Schaeffer et al 2013). Given the observed increases in tundra stream NO₃ ⁻ export from 1978 to 2006 (McClelland et al 2007, Townsend-Small et al 2011), understanding mechanisms for N availability and losses in tundra soils is important. However, estimates of nutrient leaching losses from tundra soils remain limited (Harms and Jones Jr 2012, Treat et al 2016). Rates of N mineralisation, N assimilation and immobilisation vary seasonally due to interactions with plant biomass and microbes (Schmidt et al 2002, Sorensen et al 2008, Baker et al 2023) and further complicate estimates of N leaching losses. For example, soil N losses increased significantly during the late growing season at a high-Arctic tundra site as plants senesced (Schaeffer et al 2013).

Losses of DON account for the majority of soil N leaching to streams from pristine ecosystems (Hedin et al 1995, Perakis and Hedin 2007) like Arctic ecosystems (Lafrenière and Lamoureux 2008, McNamara et al 2008). The export of dissolved OM, and therefore DON, is related to dissolved OM concentrations within the soils, dissolved OM retention within the soil, and hydrologic flux from the soil.

The export of dissolved OM is also controlled by the amount of runoff. Olefeldt and Roulet (2012) found substantially lower dissolved OM fluxes from a permafrost peatland than from other peatlands and concluded this was the result of less runoff due to presence of impermeable permafrost. Changes in the flow path of water due to permafrost thaw will alter dissolved OM and N inputs to aquatic ecosystems, resulting in predictions of increasing DON fluxes in Siberian watersheds (Frey et al 2007).

In streams, nutrient export is highly variable over time and space. The majority of nutrient export in Arctic rivers occurs during spring freshet (Holmes et al 2012, Juhls et al 2020). Holmes et al (2012) estimated dissolved N loading to be 1.26 Tg per year; roughly two-thirds (0.85 Tg) was organic N over the inorganic remainder (NO₃ ⁻ and NH₄ ⁺). Additionally, there is seasonal variability in the quantity of the dominant N form released in the Yedoma domain. As was found by Sanders et al (2022), who studied the Yedoma domain, DON is predominantly transported to the ocean in summer, while the concentrations of exported DON and NO₃ ⁻ are similar during periods with ice cover.

Besides quantity, the quality of dissolved OM is also temporally variable. OM quality differs during the spring freshet from the growing season; 20%–46% of dissolved OM lost during spring freshet is degraded relatively rapidly (Michaelson et al 1998, Holmes et al 2008), while dissolved OM exported to the Arctic ocean during the summer is slow to degrade (Holmes et al 2008). N released by dissolved and particular OM mineralisation may result in increased stream productivity or N retention (Peterson et al 2001), denitrification (Peterson et al 2001), or export to the Arctic Ocean (Frey et al 2007, Mann et al 2022, Sanders et al 2022, Ogneva et al 2023). The net effect on climate, considering direct and indirect greenhouse gas emissions from both OC and N cycles is presently difficult to estimate from lakes and rivers in the Yedoma domain, and a solid estimation for N₂O emissions from Arctic rivers is lacking until now.

Excess N liberated during thaw will be partly immobilised by being incorporated into the microbial biomass following thaw, but with continued supply of N, the likelihood of excess N availability increases. This excess N is, depending on the micro-conditions, simultaneously available for various microbial transformation processes including N₂O and molecular N (N₂) production as long as growth of vascular plants remains hampered (Palmer et al 2012, Palmer and Horn 2012, Hetz and Horn 2021, Horn and Hetz 2021).

The effect of excess N is a subject of discussion, giving rise to two distinct phenomena: the concept known as the 'N mining theory' (Lovett et al 2018) and the 'N priming effect' (Mack et al 2004). The N mining theory suggests an increase in decomposition due to N limitation, where microbes are forced to mine for additional N-substrates from OM. The theory suggests that when N is limited, N-acquiring microbes can use OC as an energy source, thereby enhancing OM decomposition and increasing respiratory CO₂ losses to acquire the needed N (Chen et al 2014, Lovett et al 2018). The N priming effect, on the other hand, can be positive or negative, entailing a boost or a decrease in overall OM decomposition through enhanced N availability.

Notably, the N-priming effect not only enhances the decomposition of OM and stimulates further N-release, but also leads to higher N availability and enhanced decomposition of OM increase OC release (Mack et al 2004, Lavoie et al 2011). For example, Mack et al (2004) showed that aboveground OC storage increased following N-addition, but was offset by increased OC decomposition—and thus soil OC losses—belowground. Similarly, Zhang et al (2023) found that a warming-induced, enhanced soil N supply stimulated soil microbial activity and amplified soil OC losses from permafrost soils in the Tibetan Plateau. These findings support the 'stoichiometric decomposition theory', which suggests that N addition is beneficial for OM decomposition.

In contrast to these findings, Berg and Matzner (1997) and Riggs et al (2015) showed that increased N availability could lead to decreased OM decomposition and thus negative priming. Riggs and Hobbie (2016) hypothesised that negative priming can be caused by three mechanisms. First, increased N availability can improve OC use efficiency leading to OC use for microbial growth rather than respiration (Weintraub and Schimel 2003, Manzoni et al 2012, Chen et al 2018). Second, N addition might lead to decreased activity of oxidative enzymes, thereby limiting decomposition of complex OC molecules such as lignin (Eisenlord et al 2013). Third, soil acidification following N addition might lead to decreased microbial biomass and hence to decreased OM decomposition rates (Mueller et al 2012).

For Alaskan Arctic and boreal soils, Lavoie et al (2011) found that N addition could have a positive as well as a negative effect on OM decomposition. They showed that a positive effect was mainly observed for organic soils, whereas negative effects were mainly found for mineral soils, from which they concluded that the direction of priming in decomposition is highly dependent on the soils' labile OM fraction. Likewise, Wologo et al (2021) found that addition of N (in the form of NH₄ ⁺ and NO₃ ⁻) and acetate, compounds abundant in permafrost including Yedoma (Mann et al 2022), triggered the breakdown of some dissolved OM compounds, but resulted in a net decrease in overall decomposition rate. Thus, depending on the OM degradability and the microbial composition, the N stock in the Yedoma domain might pose either a positive or a negative climate feedback upon thaw.

Permafrost thaw will increase the amount of bioavailable N in an N-limited environment, which could therefore boost ecosystem productivity resulting in a negative climate feedback. Walter Anthony et al (2014) suggest that N and other nutrients such as P released from thawing Yedoma deposits have supported high lacustrine primary production and OC accumulation in thermokarst lake sediments during the Late Glacial to Holocene warming. In addition, in re-vegetated (stabilised) thermokarst basins, plants will be able to access former deep Yedoma N (see figure 3).

On the one hand, as mentioned before, abrupt thaw could increase OC sequestration, while gradual thaw poses the potential for incomplete nutrient use by plants (due to temporal and spatial mismatch). On the other hand, rapid thaw processes disturb or destroy the plant cover and thereby hinder the possibility of plants to catch the initial thaw liberates nutrient pulse. Therefore, it is possible that a considerable amount of N is not used to promote OC sequestration in the short-term. As a result, the possibility to create a negative feedback to climate via enhanced plant growth is limited in these regions with on-going thermokarst activity.

While part of the excess N liberated from thawing permafrost in the Yedoma domain will be used for plant growth and transformed into microbial biomass with continued N supply as described above (figure 2), part of it may become available for microbial transformation processes such as nitrification and denitrification (Horn and Hetz 2021, Ramm et al 2022). These microbial N transformation processes may cause gaseous N losses in the form of N₂O as well as N₂. N₂O is produced mainly by denitrification but also as a side product during the first step of nitrification, the ammonia oxidation. Further, nitrification has a crucial role for denitrification via NO₃ ⁻ supply to serve as the electron acceptor in denitrification. Recently published data on high N₂O emissions and potential N₂ losses from retrogressive thaw-slumps within Yedoma deposits demonstrate that this is a likely scenario for such permafrost thaw features in the Yedoma domain (Marushchak et al 2021).

There is increasing evidence from soils across the permafrost region that, despite their general N limitation due to slow N mineralisation in cold and often humid conditions, permafrost-affected soils can be substantial sources of N₂O when three conditions are met (Repo et al 2009, Voigt et al 2017b, 2020, Marushchak et al 2021). First, a sufficient mineral N supply is a key factor controlling N₂O emissions from soils (Parton et al 1996). In this context, the C:N ratio has been linked to N₂O emissions from soils with highest emissions occurring at low to intermediate range C:N ratios, which support N mineralisation (Klemedtsson et al 2005, Liimatainen et al 2018). In the Yedoma domain, the C:N ratios ranged between 9–19 (medians of the different compartments in Strauss et al (2022), which is low enough to support N₂O emissions as shown in Marushchak et al (2021).

Second, a lack of vegetation growth increases the mineral N availability for microbes and is thus favourable for N₂O emissions. Previous studies reported high N₂O emissions in locations where vascular plants are absent or N uptake by vegetation is otherwise disturbed, e.g. by thermokarst processes (Marushchak et al 2011, Gil et al 2017, Voigt et al 2017a). This is frequently the case in the Yedoma domain with its high abundance of eroded surfaces, disrupted plant growth, and disturbed site conditions due to abrupt thaw processes.

Third, an intermediate moisture content has been shown to promote N₂O emissions from permafrost-affected soils (Voigt et al 2020). On the one hand, water-saturated conditions hinder N mineralisation and nitrification, which are aerobic processes, thereby limiting provision of mineral N species used for N₂O production. At the same time, a high water content promotes denitrification as this is an anaerobic process. If denitrification is complete under water-saturated conditions, N₂O is reduced to the inert N₂ gas, and complete denitrification may also lead to atmospheric N₂O uptake (Butterbach-Bahl et al 2013, Voigt et al 2017b, 2020). On the other hand, an intermediate moisture range allows both NO₃ ⁻ and N₂O production via aerobic nitrification as well as N₂O production via anaerobic denitrification to co-occur (Butterbach-Bahl et al 2013, Abbott and Jones 2015, Fiencke et al 2022). Thus, the magnitude of N₂O emissions following permafrost thaw will largely depend on hydrology.

Limitations in microbially available OM together with accumulation of NO₃ ⁻ associated with high nitrification rates might further enhance N₂O production during denitrification (Hetz and Horn 2021). In favourable, intermediate soil moisture conditions, a microbial community capable of intensive mineral N cycling (by nitrification and denitrification) and substantial N₂O production can be established in these mineral N rich sediments (Marushchak et al 2021). While for freshly thawed, wet Yedoma sediments, low N₂O emissions were found due to low abundance of key functional groups, particularly nitrifiers, whereas high N₂O emissions occurred after a few years of stabilisation, drying, partial re-vegetation and changes in microbial community (Marushchak et al 2021).

Based on these initial findings, we thus expect that the following areas are prime candidates for substantial N₂O release (figure 3): thaw slumps, freshly drained lake basins, and thermo-erosional valleys. If thermokarst lakes drain, unfrozen sub-lake taliks can turn into zones of intermediate soil moisture levels. Nitze et al (2020) observed widespread lake drainage in the Yedoma domain over the past decades and predicted a similar trend of more lake drainage events in the future, especially for southern margins of continuous permafrost. With improved drainage, the OM in lake sediments is exposed to enhanced mineralisation, which will boost mineral N production and the microbial N transformation processes releasing N₂O.