Introduction
Greenhouse gases (GHGs) such as carbon dioxide (CO2), methane (CH4), and nitrous oxide (N2O) can absorb infrared radiation and trap heat in the atmosphere (IPCC 2007). This greenhouse effect is critical for maintaining livable conditions for life on earth, but anthropogenic activities have caused an unprecedented increase GHG concentrations in the atmosphere, leading to global warming and climate change (IPCC 2007). Soils play an important role in climate change, as they have the potential to store carbon or to be a source of greenhouse gases to the atmosphere (Oertel and others 2016).
Riparian zones are the three dimensional zones of direct interaction between terrestrial and aquatic ecosystems (Gregory and others 1991). Riparian ecosystems are a conservation priority due to their disproportionately high value and diversity of ecological functions as well as unique biodiversity combined with high vulnerability to anthropogenic pressures (Ramey and Richardson 2017; Capon and Pettit 2018). Their unique properties, including shallow water tables, high soil organic matter content, and high soil nitrogen concentrations, create the potential for significant carbon sequestration (Gundersen and others 2010), but also to produce significant amounts of anaerobically produced CH4 and N2O to the atmosphere (Knoepp and Clinton 2009; Vidon and others 2018).
Forest harvest is a disturbance that can alter biogeochemical processes in soils (Kreutzweiser and others 2008), resulting in a change in the GHG emission rates (Lavoie and others 2013). However, soil ecosystem responses to logging are highly variable and site specific due to microsite level differences in variables such as soil properties, moisture conditions, and biological interactions (Kreutzweiser and others 2008). Therefore, the results from the many upland forest studies are likely to differ from riparian zones.
The most common method of protecting streams and riparian areas from the impacts of forest harvest is the use of riparian buffer zones, consisting of a strip of trees adjacent to the stream that is either left uncut or has limited harvesting (Richardson and others 2012). The riparian zones of headwater streams are particularly vulnerable to forest harvest, as buffer zones are often not required along these small streams (Richardson and Danehy 2007). In British Columbia, 45% of surveyed small streams did not retain a riparian buffer after forest harvest (Kuglerová and others 2020). Headwater streams are small, ecologically significant, tributaries at the most upstream ends of a stream network that make up to 80% of stream length in a given drainage network (Leopold and others 1964). Although riparian buffers are effective in reducing the impacts of forest harvest by intercepting sediments, maintaining bank stability, and providing shading among other benefits (Richardson and others 2012), their usage remains contentious in headwater systems, due to the large volume of timber that buffers remove from commercial use (Richardson and Danehy 2007).
The major drivers of spatio-temporal variation in GHG flux rates include soil temperature and soil moisture (Luo and Zhou 2006). CO2 emissions consist of the combined emissions from root respiration (autotrophic respiration) and microbial decomposition of organic matter (heterotrophic respiration), and as such, CO2 emissions are affected by substrate availability (Luo and Zhou 2006). CH4 is produced by methanogens in anaerobic conditions, and consumed by methanotrophs in aerobic conditions (Oertel and others 2016). Soil pH and substrate availability impact microbial communities contributing to CH4exchange (Serrano-Silva and others 2014). N2O is mainly produced by denitrification under anaerobic conditions, in addition to nitrification under aerobic conditions (Oertel and others 2016). Important factors influencing these processes are nitrogen deposition and fertilization, and soil pH (Dalal and Allen 2008).
Forests and trees are critical in regulating water, energy, and carbon cycles (Ellison and others 2017; Ilstedt and others 2016). Thus, deforestation can alter local scale warming, rainfall, water availability, and the emission of GHGs (Ellison and others 2017). Studies have evaluated the effects of clear-cutting on GHG fluxes from forest soils (Kähkönen and others 2002). Clear-cutting has been found to increase soil temperatures (Hashimoto and Suzuki 2004) and increase the elevation of the groundwater table (Bliss and Comerford 2002; Hotta and others 2010). These changes can be attributed to the drastic reduction in leaf area after harvesting, which reduces catchment-wide transpiration rates (Bliss and Comerford 2002) and allows for an increase in incoming sunlight, warming up the soil (Hashimoto and Suzuki 2004). Clear-cutting can increase CH4 emissions due to greater soil moisture and temperature (Wu and others 2011) and increase N2O emissions due to greater soil moisture and increased nitrogen availability as a result of increased rates of nitrogen mineralization and/or reduced competition from roots (Kellman and Kavanaugh 2008; Takakai and others 2008). Clear-cutting has inconsistent effects on CO2 emission rates, with some studies reporting a decline (Striegl and Wickland 1998), an increase (Paul-Limoges and others 2015), or no change (Kähkönen and others 2002) in emissions following harvest. Nevertheless, to our knowledge, no one has yet examined the effects of forest harvest in riparian forests alongside streams on the rates of GHG fluxes.
In this study, we evaluated how forest management practices can alter biogeochemical processes that subsequently affect GHG fluxes from riparian soils along headwater streams. The objectives were: (1) to quantify the effects of forest harvest practices on soil GHG flux rates; (2) to determine if soil temperature, soil moisture, and/or groundwater level were dominant driver(s) of gas fluxes; and (3) to determine the effects of microtopographical variation on GHG flux rates.
Methods
Site description
The study sites were located in the 5,157 ha Malcolm Knapp Research Forest (MKRF), at the foothills of the Coast Mountains, about 40 km east of Vancouver, British Columbia (49° 16’ N, 122° 34’ W) (Figure 1). The biogeoclimatic zone is Coastal Western Hemlock (Klinka and others 2005), and the dominant tree species are Western Hemlock (Tsuga heterophylla ), Douglas-fir (Pseudotsuga menziesii ), and Western Red Cedar (Thuja plicata ) (Klinka and others 2005). The forest is mostly comprised of approximately 90-year old second growth, naturally regenerated following widespread fire in 1925, and again in 1931 (Klinka and others 2005).
The climate is maritime, with slight continental influence due to the mountains and inland location (Klinka and others 2005). The primary climate (Köppen) classification is Cfb , temperate oceanic climate (Kottek and others 2006). The climate is characterized by mild temperatures, with wet, mild winters, and cool, relatively dry summers (Klinka and others 2005). Mean annual precipitation and air temperature at the Environment Canada climate station located at the Research Forest (Haney UBC RF Admin, station number 1103332) are 2131 mm and 9.7°C, respectively (data for 1962 to 2006). The annual range in mean monthly minimum and maximum temperature was 5.3°C to 13.8°C over the same period.
Glacial till and colluvium are the predominant parent materials in MKRF (Klinka 1976). In the southern portion, where our sites were located, surficial deposits include glacio-fluvial and glacio-marine deposits from Pleistocene era glaciation, overlaying compacted till or bedrock (Klinka and others 2005; Klinka 1976). The soils formed on these materials are shallow and can be expected to be coarse, acidic, and low in basic cations (Klinka and others 2005). The soils at all sites were designated a Humo-Ferric Podzol, except one site (E10B) had soil of the Organic Order (likely a Hydric Mesisol). Previously reported properties of mineral soil in undisturbed forest stands in Malcolm Knapp Research Forest included 39 Mg ha-1 total carbon content, 2.5 kg ha-1 nitrate content, and 11.5 kg ha-1 available phosphorous content (Turk and others 2008).