Water and sea-ice carbon linkages in an Arctic coastal glaciated system

Samples for carbonate system parameters, including dissolved inorganic carbon (DIC) and total alkalinity (AT), were collected in 250 mL borosilicate bottles following established protocols (Dickson et al., 2007). To preserve the samples, 50 µL of saturated mercuric chloride was added, and they were stored in the dark at 4°C until analysis. Nutrient samples were collected in 20 mL HDPE vials, frozen at −20°C, and later analysed. Sea-ice samples were collected on 20th and 23rd of April from two distinct locations using a KOVACS ice corer (Mark II coring system, inner diameter 9 cm). Cores A and B were obtained from sea ice within the northern region influenced by the glaciers Kongsbreen and Conway, and Core C was collected from the inner part and easternmost side of the study area, located near the glacial front of Kongsvegen. Each core had dimensions between 40 and 50 cm in length and 9 cm in diameter and was segmented into four to five sections from top to bottom. Each 10 cm section was thawed at 5 °C in a dark room. All samples underwent filtration for POC, DOC and CDOM after which they were stored frozen until further analysis. CDOM absorbance was measured using a dual-beam spectrophotometer Cary 4000 UV-Vis and a 100 mm cuvete, with Milli-Q water serving as a reference. Spectral measurements were collected from 200 to 800 nm at intervals of 0.5 nm. Absorbance spectra a(λ) was converted to CDOM absorption coefficients, aCDOM(λ) (m⁻¹). The absorption spectra's characteristics between 250 and 700 nm were delineated by determining the exponential spectral slope coefficient (S). An alternative approach proposed by Helms et al. (2008) was also examined, which involved calculating slopes within two different wavelength ranges (S275-295 and S350-400) and their ratio (SR) (S275–295: S350–400). The software Ocean Data View was used to create x, y plots, and section profiles employing a DIVA gridding technique to interpolate the measurements spatially between each depth profile. A Gaussian decomposition method was employed to model the CDOM absorption spectra and determine specific absorptive components. This method assumes that deviations, such as shoulders and peaks, from a typical exponential pattern in the absorption spectra are due to the significant presence of specific compounds or structures (Massicotte & Markager, 2016). The decomposition was performed over the wavelength range of 270–700 nm to capture all spectral characteristics using the Asfit software (Omanović et al., 2019). The statistical analysis of the decomposed Gaussian components was conducted in Python. To examine the relationship between parameters and the properties of CDOM, we employed Pearson's method to calculate the Pearson coefficient. Kruskal-Wallis in the software test was applied to assess differences between the stations.

碳酸系统参数样本，包括溶解无机碳（DIC）和总碱度（AT），系按照既定协议（Dickson et al., 2007）在250毫升的硼硅酸盐瓶中收集。为保存样本，添加了50微升的饱和氯化汞，并在避光条件下储存于4°C直至分析。营养样本则收集于20毫升的HDPE试管中，并在-20°C下冷冻，随后进行分析。海冰样本于4月20日和23日从两个不同地点收集，采用KOVACS冰芯钻（Mark II钻探系统，内径9厘米）。芯A和B来自受Kongsbreen和Conway冰川影响的北部海冰区域，芯C则采集自研究区域内部及最东端，靠近Kongsvegen冰川前沿。每个芯的长度介于40至50厘米之间，直径9厘米，自上而下分为四至五段。每10厘米的段落在5°C的暗室中解冻。所有样本在分析前均经过POC、DOC和CDOM的过滤处理，之后被冷冻保存。CDOM吸光度采用Cary 4000 UV-Vis双光束分光光度计和100毫米光程杯进行测量，Milli-Q水作为参照。光谱测量从200至800纳米，每隔0.5纳米进行一次。吸光度光谱a(λ)被转换为CDOM吸收系数aCDOM(λ)（m⁻¹）。在250至700纳米范围内，通过确定指数光谱斜率系数（S）来描述吸光光谱的特征。同时，也考察了Helms等（2008）提出的一种替代方法，该方法涉及计算两个不同波长范围内的斜率（S275-295和S350-400）及其比值（SR）（S275–295: S350–400）。使用Ocean Data View软件创建x, y图和剖面图，并采用DIVA网格技术对每个深度剖面之间的测量进行空间插值。采用高斯分解法对CDOM吸收光谱进行建模，并确定特定的吸收组分。此方法假定吸收光谱中偏离典型指数模式（如肩部和峰部）的偏差，是由于特定化合物或结构的大量存在（Massicotte & Markager, 2016）。分解在270至700纳米的波长范围内进行，以捕捉所有光谱特征，使用Asfit软件（Omanović et al., 2019）。分解后的高斯组分在Python中进行统计分析。为检验参数与CDOM属性之间的关系，我们采用皮尔逊方法计算皮尔逊相关系数。在软件测试中应用Kruskal-Wallis方法以评估各站点之间的差异。