A first insight into the estimation of uncertainty associated with storage and physical preparation of forest moss samples for trace element analysis

Highlights

• Moss samples were stored under three different conditions.

• After storage samples were cleaned using dry and wet techniques.

• Higher contents of elements were found in samples dry cleaned after storage.

• The level of sample storage uncertainty varied from 4.8% to 24.0%.

• Uncertainty related to preparation was in the range of 3.0–21.0%.

Abstract

To assess the level of uncertainty related to storage and physical preparation, 27 combined samples of moss species Pleurozium schreberi (Brid.) Mitt and their duplicates were collected within three forest areas. Each sample was divided into three sub-samples subjected to further treatment: D – drying (20–25 °C), F – freezing (−20 °C), and A – acclimatization (4 °C). After 7 days, all the samples were split into two sub-samples for physical preparation, dry (P1) and wet (P2) cleaning, respectively. Subsequently, the samples were milled, digested in a closed microwave system and analysed for Cd, Co, Cu, Fe, Mn, Ni, Pb and Zn using the flame (FAAS) and graphite furnace (GFAAS) atomic absorption spectrometry. In four out of eight metals (Cu, Fe, Mn, Ni), the lowest mean values were in the samples stored at −20 °C, however, in two cases (Cd, Pb), the acclimatization procedure also led to lower concentrations. Except for Mn and Zn, the higher contents of elements were found in samples that were dry cleaned after storage. The level of sample storage uncertainty (s_rstor) varied from 4.8% to 24.0%. The uncertainty related to preparation was in the range of 3.0–21.0%. Only for Co, the contribution of s_rstor uncertainty to the budget of uncertainty was lower than that from physical preparation. For the other elements, this contribution was at a similar level (Ni, Zn) or higher (Cd, Cu, Fe, Mn, Pb). In most cases, the lower s_rprep values were obtained when the samples were dry cleaned after storage, regardless of the storage conditions.

Introduction

Mosses have been used as bioindicators for at least fifty years (Rühling and Tyler, 1968). Biomonitoring surveys are conducted on a national, international or local scale (e.g., Berg and Steinnes, 1997; Fernández et al., 2000; Harmens et al., 2004, 2010; Barandovski et al., 2008; Frontasyeva and Harmens, 2014; Kłos et al., 2015, 2018), whereas mosses are used as bioindicators of organic compounds, major, trace and rare elements and isotopes (e.g., Reimann et al., 2001; Gerdol et al., 2002; Migaszewski et al., 2009, 2010; Dołęgowska and Migaszewski, 2011, 2013; Liu et al., 2012; Agnan et al., 2014; Castorina and Masi, 2015). The European Moss Survey, the most important project on the use of naturally growing mosses as bioindicators of air quality, has assumed international cooperation between European countries since the 1990s. Biomonitoring studies conducted on such a broad scale are possible thanks to standardization of procedures included in the ICP-Vegetation manual (Frontasyeva and Harmens, 2014). However, as shown by Fernández et al. (2015), some of the recommendations in this manual are unclear, some are modified by researchers due to the specificity of a sampling area, or are not taken into account. In some cases, the ICP-V manual proposes several versions of one procedure that can be used interchangeably, and sometimes does not give any recommendations (e.g., distance between sample sites or sample homogenization), which requires of researchers to make their own assumptions. In terms of moss treatment methods, the protocol says that moss samples should be cleaned immediately after sampling, or if they cannot be cleaned, they should be: (i) dried and stored at a room temperature or (ii) deep-frozen for further processing (Frontasyeva and Harmens, 2014). Another alternative method, proposed by Aboal et al. (2008), assumes that samples can be stored in the fridge at a temperature of 4–5 °C (acclimatization), provided they are cleaned within two weeks after sampling.

Drying and freezing are the methods included in the European Survey protocol, such as those proposed for the treatment of moss samples after sampling (Frontasyeva and Harmens, 2014). However, both may alter moss membrane permeability, and in consequence, change the elemental composition of samples. Aboal et al. (2008) were the first to show that storage conditions affect moss chemistry. The authors found differences in the content of selected elements in the Pseudoscleropodium purum samples, which were just after sampling dried at an ambient temperature, frozen, or acclimatized. Consequently, they expressed doubts concerning the possibility of comparing the results obtained from the analysis of samples stored under different conditions.

Changes in the permeability of moss membranes may affect further steps of the measurement process, including physical preparation, e.g., cleaning. The efficiency on this step depends on several factors, such as: (i) type and time of cleaning, (ii) amount of particles deposited on the moss surface, and (iii) condition of tissues, in particular their moisture (Aboal et al., 2008, 2011; Pérez-Llamazares et al., 2011; Spagnuolo et al., 2013; Di Palma et al., 2017). The main objective of cleaning is to remove an adhered material, but there are no well-defined rules for this process. In the literature we can find several dry (e.g., brushing, shaking, use of nitrogen or air jet) and wet (e.g., washing with tap, deionized or bidistilled water, using ultrasounds) techniques (Türkan et al., 1995; Sardans and Peñuelas, 2008; Migaszewski et al., 2009; Spagnuolo et al., 2013; Bustamante et al., 2015). As it was shown by Fernández et al. (2015) among different techniques (if given) washing is the most popular method. However, this also may change moss chemistry, which has been reflected in differences between washed and unwashed samples, or between dry and wet (e.g., rinsed) cleaned samples. This means that the use of different protocols (in relation to cleaning technique) does not give equivalent results, so these methods should not be used interchangeably. The physical sample preparation step is aside from sampling one of the most crucial steps of biomonitoring studies (Lyn et al., 2007; Lequy et al., 2016). The quality of this stage translates directly into the quality of the final results. In environmental studies, the quality of the measurement process refers mainly to sampling and chemical analysis. Therefore, the measurement uncertainty that expresses this quality can be computed as a sum of two components: sampling and analytical uncertainties (Ramsey, 1998, 2002; Ramsey and Ellison, 2007) and these authors suggested that the ideal maximum should not exceed 20% of the total uncertainty. When the sample preparation step (mainly related to cleaning) is included to these calculations then the sum of the considered components should be lower than 35% (Pasławski and Migaszewski, 2006).

Because the storage of samples is a part of the measurement process that leads to changes in moss element composition, it can be treated as another source of uncertainty. This step is carried out only if samples cannot be cleaned immediately after sampling, but due to the specificity of biomonitoring studies and a large number of samples collected, it is practically impossible to clean them immediately. So far, only a few publications have concerned the influence of moss storage conditions on their chemistry, but none of them considered this stage as a source of uncertainty. If we decided to this, the most important question is whether the uncertainty arising from this stage should be treated as a component of the measurement uncertainty or as a separate one. In 2002, Ramsey gave the answer to this question and assumed that the sample preparation uncertainty could be included to the measurement uncertainty, or could be computed separately (Ramsey, 2002).

The main objective of this study was to estimate, for the first time, the level of uncertainty arising from the storage under different conditions of moss samples taken from forest areas. We also estimated how the further treatment of moss samples (type of cleaning) affects their elemental composition and the level of uncertainty linked to the cleaning itself.