1 INTRODUCTION

Food webs in forested mountain streams usually rely on allochthonous inputs of nutrients (Vannote et al., 1980); however, these inputs are frequently seasonal. For example, in snow-pack fed streams, leaf litter and spawning salmon contributions peak in the autumn, wet atmospheric deposition peaks in winter, and soil runoff is highest during the spring melt. Determining the role of in situ nutrient sources may help to complete the picture of stream food webs when allochthonous inputs are low. One potential in situ nutrient source is biological nitrogen (N) fixation, a microbial process that converts N₂ gas into a biologically accessible form of N (i.e. ammonia). Although in-stream N fixation often has been viewed as less important than other N inputs (e.g. fertiliser runoff, marine-derived N) to streams (Howarth et al., 1988), N fixation can be the main N source to low-N streams (Carmiggelt & Horne, 1975; Grimm & Petrone, 1997). Specifically, stream N fixation could be important in the Cascade Mountains of the Pacific Northwest, which contain many sections of older volcanic rock with low N to phosphorus (P) ratios (Leland, 1995). Many Pacific Northwest streams, which formerly supported healthy salmon runs, probably receive only 6–7% of the marine-derived N and P that they historically received from salmon (Gresh et al., 2000). In addition, increased tree biomass as a consequence of fire suppression results in more stored terrestrial N, further reducing stream N inputs (Bernal et al., 2012). These conditions increase the likelihood that Cascade streams are N-depleted and therefore more heavily reliant on N fixation.

The primary groups of N-fixers in the Cascades include heterotrophic microbes in benthic sediments and aquatic cyanobacteria, predominantly Nostoc paramelioides (family Nostocaceae), which is filamentous and forms colonies (Dodds et al., 1995). Nostoc paramelioides colonies are present in two types: (1) those that form a symbiotic relationship with the chironomid (midge) larva Cricotopus nostocicola or fuscata (Brock, 1960), hereafter referred to as “cyano-midge”, and (2) those without a midge symbiont, hereafter referred to as “cyano-only.” The midge larva tunnels into the Nostoc colony and uses the colony as food until pupation (Brock, 1960). In return for food and protection from predators, the midge benefits the Nostoc colony by attaching it firmly to the rocks, changing its shape to improve gas exchange, and increasing Nostoc dispersal via the spread of reproductive filaments during midge emergence (Dodds & Marra, 1989). The relationship between N. paramelioides and the midge may impact N fixation rates by changing the shape and surface area of the colony to increase diffusion and add stability, especially in currents faster than 10 cm s⁻¹ (Dodds, 1989), conditions which are common in high-gradient Cascade streams.

In addition to possible impacts from biological factors, many physical and chemical factors influence N fixation, which is a high-energy process, theoretically requiring 16 ATP to fix one N₂ molecule (Kim & Rees, 1994). It is catalysed by the nitrogenase enzyme, which contains P and micronutrients. For autotrophic cyanobacteria, N fixation tends to increase with light (Berrendero et al., 2016; Carmiggelt & Horne, 1975; Grimm & Petrone, 1997), P (Marcarelli & Wurtsbaugh, 2006, 2007, Kunza & Hall, 2013, but see Scott et al., 2009), temperature (Marcarelli & Wurtsbaugh, 2006; Welter et al., 2015) and substrate stability (Marcarelli & Wurtsbaugh, 2009). Dissolved inorganic N (DIN) frequently inhibits N fixation in cyanobacteria, because they can meet their nutritional needs with less energy (Eberhard et al., 2018; Hiatt et al., 2017; Kunza & Hall, 2013, 2014; Marcarelli & Wurtsbaugh, 2006, 2007; Scott et al., 2009). Likewise, heterotrophic N fixation rates increase with carbon (C; energy) availability, such as that provided by leaf litter (Tam et al., 1981) and fine sediment (Francis et al., 1985). Heterotrophic N fixation also increases with P (Romero et al., 2012) and often is inhibited by DIN (Caton et al., 2018; Eberhard et al., 2018, but see Knapp, 2012). Although cyanobacterial and heterotrophic N fixation are influenced by similar physicochemical drivers, the midge symbiosis may cause distinct patterns in cyano-only and cyano-midge fixation rates.

Drivers of N fixation are well-established in general, but they have not been thoroughly explored in streams (Marcarelli et al., 2008). In particular, the relative contributions of different taxa have not been examined. Notably, very few studies have measured N fixation rates in sediment heterotrophs, which recently have been identified as important N-fixers in marine environments (Aoki & McGlathery, 2019; Newell et al., 2016; Rao & Charette, 2012), suggesting that they also could be important in streams.

We measured N fixation rates, along with possible explanatory variables, in seven forested streams in the Cascade Mountains, Washington, USA, during the summer and autumn of 2019. Our objectives were to (a) determine the relative importance of the major groups of N-fixers: cyano-midge, cyano-only and sediment microbes, and (b) establish rates and physicochemical drivers of N fixation. We hypothesised that light would be the primary predictor of cyanobacteria N fixation. Likewise, we predicted that cyanobacteria would fix more N than sediment microbes in streams with higher light availability, whereas sediment heterotrophs would fix more N in shaded streams. Finally, we hypothesised that the presence of a midge symbiont would increase cyanobacteria N fixation rates in our high-gradient streams.

中文翻译：

---

森林覆盖的山间溪流中的固氮率：沉积物微生物比以前认为的更重要吗？

1 简介

森林覆盖的山间溪流中的食物网通常依赖于外来的营养物质输入（Vannote 等人，  1980 年）；然而，这些投入通常是季节性的。例如，在积雪喂养的溪流中，落叶和产卵鲑鱼的贡献在秋季达到高峰，潮湿的大气沉降在冬季达到高峰，而土壤径流在春季融化期间最高。当外来输入量较低时，确定原位营养源的作用可能有助于完成河流食物网的图景。一种潜在的原位营养来源是生物固氮 (N)，这是一种将 N _{2转化为微生物的过程}气体转化为可生物利用的 N 形式（即氨）。尽管河流中的 N 固定通常被认为不如河流的其他 N 输入（例如肥料径流、海洋衍生的 N）重要（Howarth 等人，  1988 年），但 N 固定可能是低 N 的主要 N 来源流（Carmiggelt 和 Horne，  1975 年；Grimm 和 Petrone，  1997 年）。具体来说，在太平洋西北部的喀斯喀特山脉中，河流固氮可能很重要，这些山脉包含许多具有低 N 与磷 (P) 比率的古老火山岩部分（Leland，  1995 年）。许多太平洋西北部溪流，以前支持健康的鲑鱼流，可能只获得了它们历史上从鲑鱼中获得的海洋衍生 N 和 P 的 6-7%（Gresh 等人， 2000 年）。此外，由于灭火而增加的树木生物量导致更多的陆地氮储存，进一步减少了流 N 输入（Bernal 等人，  2012 年）。这些条件增加了级联流 N 耗尽的可能性，因此更严重地依赖于 N 固定。

级联中的主要固氮菌群包括底栖沉积物中的异养微生物和水生蓝藻，主要是拟念珠菌（念珠菌科），它是丝状的并形成菌落（Dodds 等人，  1995 年）。Nostoc paramelioides菌落以两种类型存在：（1）与摇蚊（蠓）幼虫Cricotopus nostocicola或fuscata（Brock，  1960 年）形成共生关系的那些，以下称为“氰基蠓”，以及（2）那些没有蠓共生体，以下称为“仅氰基”。蠓幼虫钻入发菜菌落并将菌落用作食物直到化蛹（Brock，  1960）。作为食物和保护免受捕食者侵害的回报，蠓通过牢固地附着在岩石上，改变其形状以改善气体交换，并通过在蠓出现期间繁殖细丝的传播增加发菜的传播， 从而使发菜群体受益（Dodds & Marra， 1989）。N. paramelioides和蠓之间的关系可能会通过改变菌落的形状和表面积以增加扩散和增加稳定性来影响 N 固定率，尤其是在水流速度快于 10 cm s ^-1的情况下（Dodds，  1989 年），这些条件是在高梯度级联流中很常见。

除了生物因素的可能影响外，许多物理和化学因素都会影响N的固定，这是一个高能过程，理论上需要16个ATP才能固定一个N ₂分子（Kim & Rees，  1994）。它由含有磷和微量营养素的固氮酶催化。对于自养蓝藻，N 固定趋向于随光照增加（Berrendero 等人，  2016 年；Carmiggelt 和 Horne，  1975 年；Grimm 和 Petrone，  1997 年），P（Marcarelli 和 Wurtsbaugh，  2006 年，2007 年，Kunza 和 Hall，  2013 年，但参见Scott 等人，  2009 年），温度（Marcarelli & Wurtsbaugh，  2006 年）; Welter et al.,  2015 ) 和底物稳定性 (Marcarelli & Wurtsbaugh,  2009 )。溶解的无机氮 (DIN) 经常抑制蓝藻中的 N 固定，因为它们可以用更少的能量满足其营养需求（Eberhard 等人，  2018 年；Hiatt 等人，  2017 年；Kunza 和 Hall，  2013 年，2014 年；Marcarelli 和 Wurtsbaugh，  2006 年，2007 年；斯科特等人，  2009 年）。同样，异养固氮率随着碳（C；能量）可用性的增加而增加，例如落叶层（Tam 等人，  1981 年）和细沉积物（Francis 等人，  1985年）提供的固氮率）。异养固氮也随着 P 的增加而增加（Romero 等人，  2012 年），并且经常被 DIN 抑制（Caton 等人，  2018 年；Eberhard 等人，  2018 年，但参见 Knapp，  2012 年）。尽管蓝藻和异养 N 固定受到类似物理化学驱动因素的影响，但蠓共生可能导致仅氰基和氰基蠓固定率的不同模式。

N 固定的驱动因素总体上是公认的，但它们还没有在溪流中得到彻底的探索（Marcarelli 等人，  2008 年）。特别是，尚未检查不同分类群的相对贡献。值得注意的是，很少有研究测量沉积物异养生物的固氮率，最近已被确定为海洋环境中重要的固氮剂（Aoki & McGlathery,  2019 ; Newell et al.,  2016 ; Rao & Charette,  2012），这表明它们在流中也可能很重要。

我们在 2019 年夏季和秋季期间测量了美国华盛顿喀斯喀特山脉的七个森林溪流中的 N 固定率以及可能的解释变量。我们的目标是 (a) 确定 N 主要组的相对重要性-固定器：氰基蚊、仅氰基和沉积物微生物，以及 (b) 确定 N 固定的速率和物理化学驱动因素。我们假设光将是蓝藻固氮的主要预测因子。同样，我们预测蓝藻将比具有较高光照可用性的河流中的沉积物微生物固定更多的 N，而沉积物异养生物将在阴影河流中固定更多的 N。最后，我们假设蠓共生体的存在会增加我们的高梯度流中的蓝藻固氮率。