The water samples were collected in the ultra- to oligotrophic Petrified Forest Creek and oligo- to mesotrophic Ornithologist Creek, in the vicinity of the H. Arctowski Polish station on King George Island, South Shetlands, in maritime Antarctica. These creeks were visually morphologically distinct and were selected for the study of in situ periphyton relative primary production by Elster & Komarek (2003).

Both creeks originate in one (Ornithologist Creek) or several (Petrified Forest Creek) snow patches and flow into Admiralty Bay. Oligotrophic Petrified Forest Creek has a more granulated bottom, while Ornithologist Creek is in a more open valley, has a gradual stream slope, is situated at a lower elevation and is wider. The lower reaches of Ornithologist Creek flow through a penguin rookery. Four sampling sites of the oligotrophic Petrified Forest Creek were located at 85–106 m (Reach A), 245–263 m (Reach B), 461–476 m (Reach C) and 878–885 m (Reach D) from the seashore. Three sampling sites of the oligo-to mesotrophic Ornithologist Creek were located at 68–98 m (Reach A; mesotrophic due to vicinity of the penguin rookery; the nutrient enrichment is discussed in Elster & Komarek 2003 in detail), 338–353 m (Reach B, oligotrophic) and 695–713 m (Reach C, oligotrophic) from the seashore. Their detailed geomorphological characteristics and relative primary production field data are described elsewhere (Elster & Komarek 2003).

Water samples were collected during the summer of 1996–97, at 10-day intervals from 21 December 1996 to 10 March 1997. At each sampling point, a 250-ml sample volume was collected at a depth of less than 10 cm from the surface, filtered through GF/C filters (Whatman, Kent, UK) pre-washed by stream water, and frozen for transport. All chemical analyses and measurements of AGP and APPR were performed at the Institute of Botany in Třebon, Czech Republic. The frozen samples were kept at −25°C and were completely thawed before analyses or inoculation. The chemical analyses and AGP/APPR evaluation were completed within several weeks after arriving from Antarctica (Elster & Komarek 2003).

Phosphorus [as dissolved reactive phosphorus (DRP)] and nitrogen [as dissolved inorganic nitrogen (DIN)] were analyzed using flow injection analyses (Tecator, Höganäs, Sweden; Růžička & Hansen 1981). The DRP concentration was estimated from the reaction of P O 4 3 - −P with ammonium molybdate, and reduction by the reaction of stannous chloride to phosphomolybdenum blue (Tecator application note ASN 60/83; detection limit 5 µg l⁻¹; Proctor & Hood 1954). The DIN concentration was calculated as the sum of concentrations of N H 4 + −N and N O 3 - −N. The N H 4 + −N concentration was determined by the gas diffusion method (Tecator application note ASN 50-0187; detection limit 10 µg l⁻¹; Karlberg & Twengstrom 1983) and N O 3 - −N concentration by reaction with sulphonamide (Tecator application note ASN 62-01/83; detection limit 3 µg l⁻¹).

From the total of 137 water samples—56 from Ornithologist Creek and 81 from Petrified Forest Creek—collected in 1996–97 (Elster & Komarek 2003), 30 samples were selected for the AGP and APPR analyses according to their nutrient content and sampling date. Ten samples originated from the oligotrophic Petrified Forest Creek, 10 from the oligotrophic reach of Ornithologist Creek and 10 from the lower mesotrophic reach of Ornithologist Creek (which was influenced by penguin excrement). The examined samples thus cover the full range of nutrient concentrations available in the studied streams.

The chlorococcal microalgae of the genus Stichococcus (Trebouxiophyceae, Chlorophyta) were selected for the AGP and APPR analyses, because of their wide ecological tolerance. The strains Stichococcus bacillaris Elster, 1998/28/51 (the polar strain isolated in Arctic periglacial soil in the vicinity of Ny-Ålesund, Svalbard) and Stichococcus bacillaris Hindák, 1984/15 (the temperate strain isolated in Lake Hafnersee, Austria) were used. Both strains were obtained from the Culture Collection of the Algal Laboratory (CCALA) in Třebon, and pre-cultivated at 5–7°C for the polar strain and at 18–20°C for the temperate strain, under an irradiance of 50 µmol m⁻² s⁻¹ in Z-medium (Staub 1961).

The AGP and algal primary production rate (APPR) methodology (Lukavský 1992) was modified for the water samples of polar streams. The algae were cultivated in microplates (96 wells, flat bottoms) that were thoroughly washed in 1% (v/v) HCl and later in distilled water. The surrounding wells of the plate were filled with distilled water in order to decrease evaporation. The inner wells were filled with two samples, each sample in five columns with six wells. Each well contained a 200-µl sample. The columns were: the non-treated creek water sample (sample), creek water samples with varying enrichment levels (1) 1000 µg N O 3 - −N l⁻¹ (+N), (2) 50 µg P O 4 3 - −P (+P), (3) a combination of 1000 µg N O 3 - −N l⁻¹ and 50 µg P O 4 3 - −P l⁻¹ (+NP) and (4) 1000 µg Na₂-EDTA l^−1. The additions of N and P were used to detect possible nitrogen and/or phosphorus limitations, and the addition of EDTA was used for the detection of any possible heavy metal inhibition. For the growth control, used as test validation, the strains were inoculated in a plate with the dilution series of the Z-medium (undiluted medium 1), 0.3, 0.1, 0.03, 0.01, 0.003, 0.001, 0.0003 and 0.0001 dilution factors, plus distilled water (0) as the negative control, each dilution in one column of six wells. As with the water samples, the volume of the medium was 200 µl in each well. All filled microplates were pre-sterilized by UV-C radiation for 2 h, and covered with lids and left in the dark overnight.

The pre-cultivated dense microalgal suspension of Stichococcus cells was washed by centrifugation in sterile 10-ml tubes at 315 g for 20 min to remove the medium. After the first centrifugation of the cell suspension in the medium, the cells were re-suspended in 0.5% (w/v) KCl in distilled water. This procedure was repeated twice. After washing, the cell density in 0.5% KCl was measured by cell counting in Bürker's chamber, and the inoculum volume was calculated to get the initial concentration of 10⁵ cells ml⁻¹ to reach the lower detection limit of the plate reader (Kvíderová 2010). After inoculation, the plates were covered with transparent polyethylene food foil in order to reduce evaporation during cultivation.

The cultivation lasted until the stationary phase of the growth curve was reached in each treatment, which usually took 30–40 days. The polar strain was grown at 5–7°C, and the temperate strain at 23–25°C, both in a closed cultivation unit (Labio, Prague, Czech Republic). The polar strain growth temperature reflects natural Arctic conditions, while 25°C has been recommended as the standard temperature for the AGP/APPR tests. An irradiance of 15 µmol m^–2 s⁻¹ was used during the first two days of the experiment to acclimate the algae to the new cultivation conditions; thereafter, the irradiance was increased to 50 µmol m⁻² s⁻¹. The irradiance was measured by a PU-550 digital luxmeter (Metra Blansko, Bansko, Czech Republic) with the sensor modified for the measurement of photosynthetically active radiation in the range 400–700 nm. To prevent possible carbon depletion, the cultivation chamber of the unit was enriched with a 2% (v/v) CO₂+air mixture and the microplates were shaken by a linear shaker at 50 rpm. The optical density at 750 nm (A₇₅₀) was measured every second or third day by an iEMS automatic plate reader (LabSystems, Vantaa, Finland).

Values of A₇₅₀ were converted to dry weight (DW, mg ml⁻¹) according to conversion equations. See Kvíderová & Lukavský (2003) and Kvíderová (2010) for the conversion equation estimation procedure.

The equation for temperate Stichococcus bacillaris Hindák, 1984/15 is DW ( mg m l - 1 ) = A 750 - 0.0405 1.1116 ( r 2 = 0.9895 )

The equation for polar Stichococcus bacillaris Elster, 1998/28/51 is DW ( mg m l - 1 ) = A 750 - 0.0450 0.717 ( r 2 = 0.9893 )

APG was determined as the highest observed DW content in the given sample at both 5°C (AGP₅) and 25°C (AGP₂₅). APPR was calculated as the slope of the linear regression of the lnDW versus time during the exponential phase of the growth curve (Kvíderová & Henley 2005) in each well of the microplate at both 5°C (APPR₅) and 25°C (APPR₂₅).

For a rough estimation of the gross primary production rate (µg C cm⁻² h⁻¹), the following assumptions were considered for the carbon content (µg C cm⁻²) calculation. The well bottom area of the microplate was colonized by all microalgal cells in the well and the carbon content of the DW was 50% (Padisák 2003).

The tested null hypothesis stated that there were no effects of nutrient or EDTA additions on the standardized AGP or APPR values. The null hypothesis also stated that there were no differences in AGP and/or APPR between the strains in the growth control. Before processing the data, the measured A₇₅₀ values were subjected to K-criterion statistics in order to exclude outliers; the K-criterion was performed for n = 6 and significance level of 0.05 (Likeš & Laga 1978). The normality of the data was also tested and no transformation was required. The differences in AGP and/or APPR between the strains in the growth control were tested by t-test. The difference in AGP/APPR in individual control strains was tested by one-way ANOVA. Nutrient limitation (effect of nutrient addition) was tested by two-way ANOVA, and the heavy metal inhibition tests by one-way ANOVA, using Statistica 9.0 software (StatSoft, Tulsa, OK, USA). The comparison of strain-specific nutrient requirements was evaluated by a t-test for each dilution. The results were considered significant if the p-value was lower than 0.05.

The range of nutrient concentrations in the water samples studied corresponds to the cultivation medium dilutions in the range from 0.0001 to 0.3 (Fig. 1). The DRP and DIN concentrations in the samples approximately correspond to media dilutions in the ranges 0.0003–0.1 (5–283 µg DRP l⁻¹) and 0.0001–0.003 (20–185.9 µg DIN l⁻¹) for the oligotrophic Petrified Forest Creek; 0.001–0.1 (10.5–195 µg DRP l⁻¹) and 0.0003–0.01 (58.2–574 µg DIN l⁻¹) for the oligotrophic reach of Ornithologist Creek, and finally 0.001–0.03 (7.45–124 µg DRP l⁻¹), and 0.003–0.3 (583.3–9663 µg DIN l⁻¹) for the reach of Ornithologist Creek influenced by penguin excrements. Concentrations of N H 4 + −N in the water samples of all three locations were always higher than in the Z-medium, that is, above 1.7 µg N H 4 + −N l⁻¹. N H 4 + −N concentrations ranged from 7.78 to 44.8 µg N H 4 + −N l⁻¹ in oligotrophic Petrified Forest Creek, from 10 to 66.6 µg N H 4 + −N l⁻¹ for the oligotrophic reach of Ornithologist Creek, and from 10 to 571 µg N H 4 + −N l⁻¹ for the mesotrophic reach of Ornithologist Creek influenced by penguin excrement. Concentrations of N O 3 - −N in the samples corresponds to media dilutions in the range 0.0001–0.003 (10–174 µg N O 3 - −N l⁻¹) for the oligotrophic Petrified Forest Creek, 0.0003–0.01 (37.3–564 µg N O 3 - −N l⁻¹) for the oligotrophic reach of Ornithologist Creek, and 0.003–0.3 (545–9560 µg N O 3 - −N l⁻¹) for the mesotrophic reach of Ornithologist Creek influenced by penguin excrement. The Z-medium molar N:P ratio of 33.4 was higher than in the oligotrophic Petrified Forest Creek (0.17–19.1). The N:P ratio values in the oligotrophic reach of Ornithologist Creek ranged from 3.28 to 35.4 and were in the range of lower to slightly exceeding the N:P ratio of the Z-medium, indicating a possible nitrogen limitation in the oligotrophic Petrified Forest Creek and in the oligotrophic reaches of Ornithologist Creek. However, the Z-medium N:P ratio was lower than in the mesotrophic reach of Ornithologist Creek (101–712), indicating a possible phosphorus limitation.

Data from the growth control plates revealed lower nutrient requirements of the polar strain (Fig. 2, Table 1). The AGP of the polar strain was zero (or close to zero) in dilutions from 0.01 to 0.0001 and in distilled water, and increased rapidly for Z-medium dilutions above 0.01. The maximum AGP at medium dilution of 0.3 was followed by a slight decline in the undiluted Z-medium, indicating possible nutrient over-saturation. The temperate strain AGP concentration was very low in Z-medium dilutions up to 0.03, and then the AGP rose continuously. No nutrient saturation or even oversaturation was observed in the temperate strain in undiluted Z-medium. The AGP of both strains varied at different nutrient concentrations, the polar strain started to grow with lower nutrients, that is, at higher Z-medium dilutions, and produced significantly higher biomass in Z-media dilutions in the range 0.03–0.3 compared to the temperate strain. In undiluted Z-medium, the AGP of the polar strain was lower than the AGP of the temperate strain (Fig. 2, Table 1).

Similar responses of both strains to nutrient availability were observed for measurements of APPR (Fig. 2, Table 1). The APPR of the polar strain was negligible in Z-medium dilution of 0.01 or less. The APPR suddenly reached a maximum at Z-medium dilution of 0.03 and remained stable even in undiluted Z-medium. The APPR of the temperate strain was low at Z-medium dilutions up to 0.003 and increased continuously reaching its maximum at a Z-medium dilution of 0.3. At very high dilutions (0.01–0.0001), the polar strain had a lower APPR than the temperate one. In the dilution range from 0.3 to 0.03, in which the AGP of the polar strain was significantly higher, the APPRs of the polar and temperate strains were comparable (Fig. 2, Table 1).

The AGP₅ (polar strain) of the non-treated samples was higher than AGP₂₅ (temperate strain) by 350% in the oligotrophic Petrified Forest Creek (one-way ANOVA, n=120, F=115.8, p<0.001; Table 2, Fig. 3), by 467% the oligotrophic reach of Ornithologist Creek (one-way ANOVA, n=120, F=172.7, p<0.001; Table 2 , Fig. 3), and by 516% in the mesotrophic reach of Ornithologist Creek (one-way ANOVA, n=120, F=221.3, p<0.001; Table 2 , Fig. 3).

At low temperatures (5–7°C), the AGP₅ of the mesotrophic reach of the Ornithologist Creek was higher by 48 and 40%, than AGP₅ of oligotrophic Petrified Forest Creek and of the oligotrophic reach of Ornithologist Creek, respectively (one-way ANOVA, n=180, F=78.1, p<0.001; Fig. 3 , Table 2). At moderate temperature (25°C), the AGP₂₅ of the mesotrophic reach of the Ornithologist Creek was 23% higher and 33% lower than AGP₂₅ of the oligotrophic Petrified Forest Creek and the oligotrophic reach of Ornithologist Creek, respectively (one-way ANOVA, n=180, F=21.2, p<0.001; Fig. 3 , Table 2).

The N enrichment increased AGP₅ (polar strain) by 63% in the oligotrophic Petrified Forest Creek and by 58% in the oligotrophic reach of Ornithologist Creek compared to AGP₅ without any nutrient enrichment (Fig. 3 , Table 2). Similarly, the AGP₂₅ (temperate strain) also rose significantly after N enrichment by 35 and 66% compared to AGP₂₅ without any nutrient enrichment in the two creeks specified above (Fig. 3 , Table 2). The N addition did not affect the AGP₅ or AGP₂₅ in the mesotrophic reach of Ornithologist Creek (Fig. 3 , Table 2).

The slightly positive effect of additional P, an increase of 6% as compared to AGP without any nutrient enrichment, was observed only in the oligotrophic Petrified Forest samples at 5°C. However, at 25°C, the stimulation effect of the P addition was not confirmed (Fig. 3 , Table 2). The P enrichment did not affect the AGP₅ and AGP₂₅ in the oligotrophic and mesotrophic reaches of Ornithologist Creek (Fig. 3 , Table 2).

The N + P treatment resulted only in significantly higher AGP (93%) as compared to AGP without any nutrient enrichment at 25°C in the oligotrophic Petrified Forest Creek (Fig. 3 , Table 2). The addition of N + P did not increase the AGP₅ and AGP₂₅ of the oligotrophic samples from Ornithologist Creek or of the mesotrophic reach of Ornithologist Creek.

Despite the higher AGP₅ in contrast to AGP₂₅, the APPR₅ (polar strain) of the non-enriched samples was lower than APPR₂₅ (temperate strain) by 17% in oligotrophic Petrified Forest Creek (one-way ANOVA; n=120; F=4.85; P=0.029; Table 2 and Fig. 3) and by 10% in the mesotrophic reaches of Ornithologist Creek (one-way ANOVA; n=118; F=78,0; p<0.001; Table 2 , Fig. 3), with the exception of APPR₅, which was higher by 28% in the oligotrophic reach of Ornithologist Creek (one-way ANOVA; n=120; F=6.73; p=0.011; Table 2 , Fig. 3).

Similar to AGP₅, APPR₅ was stimulated by 19 and 22% in the oligotrophic Petrified Forest Creek and the oligotrophic reach of Ornithologist Creek, respectively (Fig. 3 , Table 2), only by N enrichment. However, at 25°C, only the APPR₂₅ of the oligotrophic reach of Ornithologist Creek was significantly increased (by 49%, Fig. 3 , Table 2) and no effect was detected in the oligotrophic Petrified Forest Creek. In the mesotrophic reach of Oligotrophic Creek, the N addition did not affect the APPR₅ or APPR₂₅ (Fig. 3 , Table 2).

In contrast to AGP₅, no effect of P or N + P addition on APPR₅ and APPR₂₅ was observed either in the oligotrophic Petrified Forest Creek, or in the oligotrophic and mesotrophic part of Ornithologist Creek (Fig. 3 , Table 2). The different response pattern of AGP and APPR at 5 and 25°C was probably caused by different nutrient requirements of the experimental strains mentioned above.

No statistically significant heavy metal inhibition of AGP and APPR was observed with either strain or at either temperature (Table 2 , Fig. 4).

Correlation analysis confirmed that AGP was a good predictor of water trophic status, expressed as the molar sum of inorganic nitrogen in the form of N O 3 - −N and N H 4 + −N and inorganic phosphorus in the form of PO₄ ^{3 −} −P, at both temperatures (r ²=0.788, p<0.001 for AGP₅; r ²=0.647, p<0.001 for AGP₂₅). No such statistically significant correlations were found for APPR₅ and APPR₂₅.

The recalculated gross primary production rates, based on given assumptions and AGP₅ and AGP₂₅ values, corresponded to 0.003±0.002 and 0.003±0.002 µg C cm⁻² h⁻¹ in oligotrophic Petrified Forest Creek, 0.003±0.001 and 0.006±0.002 µg C cm⁻² h⁻¹ in the oligotrophic reach of Ornithologist Creek and 0.004±0.001 and 0.010±0.004 µg C cm⁻² h⁻¹ in the mesotrophic reach of Ornithologist Creek, respectively (mean + standard deviation, n=6).

Comparisons of standardized AGP and/or APPR with field primary production measurements will always be difficult, considering that the methods for gross primary production measurement used in polar hydro-terrestrial environments cannot easily be applied in laboratory experiments. There are several reasons why data from laboratory experiments and field measurements differ. First of all, our laboratory tests (AGP) were performed as batch cultivations under well-defined, stable conditions using defined algal strains, while in the Antarctic diverse microalgal/cyanobacterial stream community is adapted to local environmental conditions and the stream could be considered as continuous cultivation. Second, recalculation of the DW-to-carbon content can be inaccurate due to various assumptions; for example, for recalculation of DW to organic carbon we used a value of 50% (Padisák 2003). However, organic carbon content in the DW of Stichococcus strains could be higher and differ at different growth stages. The recalculated values of the gross primary production rates from our laboratory data are two or three orders of magnitude lower than previously measured field data. The gross periphyton primary production rate in polar streams (measured either as ash-free DW, oxygen production, or as ¹⁴C uptake) was in the range 0.39–4.53 µg C cm⁻² h⁻¹ (Vincent & Howard-Williams 1986; Hawes 1993; Hawes & Howard-Williams 1998; Elster & Komarek 2003). In addition, Elster & Komarek (2003) and Howard-Williams & Vincent (1989) also showed that benthic stream communities in Antarctica, either in the ponds of the McMurdo Ice Shelf or in the snow-melt streams on King George Island, were extremely rich in non-chlorophyll organic carbon. Even if it is considered that the phytoplankton primary production in polar lakes, expressed as the photosynthesis rate, is generally low compared to other cold waters (below 0.16–1.12 g C g⁻¹ Chl a h⁻¹; Markager et al. 1999). In addition, the phytoplankton biomass content is about one order of magnitude lower compared to the benthic community (Bonilla et al. 2005), therefore these field data are still higher than our laboratory measurements.

In spite of the differences mentioned above, comparison of the patterns observed in the field and the laboratory should reveal important information on ecosystem processes. We found such a discrepancy in the experimental data sets. While the relative primary productivity was lower in the mesotrophic reach of Ornithologist Creek in situ (Elster & Komarek 2003), its AGP values were higher in laboratory experiments, indicating the presence of some processes that impede microalgal/cyanobacterial growth in the field.

Although standardized laboratory tests do not include all environmental factors affecting primary production in the field, the introduced microplate method can essentially still be used for routine water trophic status monitoring and/or mineral nutrients limitation testing using AGP as the main evaluation parameter. In central Europe, various authors evaluated AGP in order to quantify the amounts of inorganic and organic compounds in water, ranging from katarobic to hypertrophic (Miller et al. 1978; Sládecek 1979; Marvan et al. 1981; Žáková 1986). Our data from Ornithologist Creek correspond to both oligotrophic (upper reach of the creek) and mesotrophic (lower reach of the creek influenced by penguin excrement) types, whereas water from the oligotrophic Petrified Forest Creek has an ultra- to oligotrophic character.

According to our results, AGP seems to be a good proxy for water trophic status even in a modified microassay due to a good correlation between the nutrient content and AGP at both experimental temperatures. In contrast to AGP, APPR does not seem to be a good predictor of water trophic status, especially in oligotrophic waters. Even though the APPR values follow those of the AGP, the pattern is not statistically significant. This is probably due to the spectrophotometric measurements being the basis of APPR estimates. The tested water samples had a very low nutrient content; the observed A₇₅₀ values did not usually exceed 0.1 in the stationary phase. Experimental errors, which are approximately 10%, could hide the effects of individual treatments in the exponential growth phase. The significant differences in AGP and APPR at dilutions of 0.01 and lower were probably caused by the use of different brands of microplates, in terms of their absorbance at 750 nm, because AGP is estimated from A₇₅₀ values. The difference in A₇₅₀ could even reach 0.005–0.01, corresponding to 0.007 –0.013 mg DW ml⁻¹ in the polar strain and 0.004–0.009 mg DW ml⁻¹ in the temperate strain (Kviderova, unpubl. data). Thus, the different brands of microplates can significantly affect the measurements in a much diluted medium. However, the data can be corrected by subtracting the values measured in filtered distilled water before data processing. Other methods for biomass content evaluation, for example chlorophyll fluorescence measurement, should increase the sensitivity of the AGP/APPR method. The application of fluorescence techniques can increase the sensitivity of the bioassay methods (Kvíderová 2010).

Although the uptake kinetics and the exact individual nutrient requirements for both studied strains are not known in detail, the dilution series method revealed that the polar Stichococcus strain produces more biomass in lower nutrient concentrations than the temperate one. The control samples (Z-medium dilution series) revealed that the AGP of the polar strain had a higher value, in comparison with the temperate one, in Z-medium dilution range from 0.03 to 0.3, that is, nutrient concentrations similar to water samples. In addition, the AGP₅ of the tested Antarctic stream water samples was significantly higher than AGP₂₅. Since APPR₅ and APPR₂₅ are comparable at Z-medium dilutions of 0.3 to 0.03, it seems that Arctic S. bacillaris is better adapted to low nutrient conditions, possibly due to higher nutrient uptake efficiency and/or better utilization of mineral nutrients. The adaptation of nutrients uptake or nutrients utilization may reflect the conditions at the original localities. The polar strain was isolated from newly deglaciated soil (in the vicinity of Ny-Ålesund in Svalbard) in a nutrient-poor and low-temperature environment. In similar ecosystems, nutrient concentrations range from 0.2 to 0.84 g N_tot kg⁻¹ in soil, and up to 15 mg P_tot kg⁻¹ in soil under a cyanobacterial mat (Liengen & Olsen 1997), and can reach values of up to 52.5 ppm NO₃ and 0.14 ppm P₂O₅ (Minami et al. 1996). However, the temperate strain was isolated from slightly mesotrophic Lake Hafnersee, Austria where the N H 4 + −N concentrations ranged from 231 to 2550 µg l⁻¹ at a depth of 10 m, and the N O 3 - −N concentration ranged from 39 to 832 µg l⁻¹ at a depth range of 0–6 m, while phosphorus concentrations (as P_tot) ranged from 8 to 30 µg l⁻¹ at a depth range of 0–10 m in 2006 (Schulz et al. 2007). Since the negative effect of elevated nitrogen on microalgal growth was proven in polar soil Chlorella-like species (Shukla et al. 2011), the inhibition of polar Stichococcus strain growth in high nutrient content seems to be possible.

The AGP₅ and AGP₂₅ of the ultra- to oligotrophic Petrified Forest Creek and the oligotrophic reach of Ornithologist Creek were significantly lower than the mesotrophic waters of Ornithologist Creek influenced by the penguin rookery, indicating lower nutrient content in the ultra- to oligotrophic samples. Mineral nutrient content in the water samples was the principal factor influencing primary production of the Antarctic freshwater ecosystem with N and P being the most important limiting nutrients (Elster & Komarek 2003). It has been shown that Antarctic hydro-terrestrial ecosystems (lotic—lakes and pools, and lentic—streams) differ in nutrient (N and P) limitations. Laybourn-Parry (2003) concluded that most Antarctic lake phytoplankton systems are limited by phosphorus, while nutrient limitation has only a minor effect on periphyton growth in benthic communities (Hawes & Howard-Williams 1998; Elster & Komarek 2003). Similar results were also observed in High Arctic lakes; nutrient limitation was proven only for phytoplankton, probably promoting the growth of chlorophyte and chromophyte microalgae, while the benthic community seems to be insensitive to the addition of nutrients (Bonilla et al. 2005). Several studies of Antarctic streams concluded that mineral nutrient limitation appears to have a minor effect on periphyton growth in a lotic ecosystem, indicating that the periphyton does not likely affect nutrient limitation in both lotic and lentic habitats (Vincent & Howard-Williams 1986; Howard-Williams & Vincent 1989; Hawes & Brazier 1991; Vincent et al. 1993; Hawes & Howard-Williams 1998; Elster & Komarek 2003).

The N and P concentrations and their ratio define the AGP, APPR and species composition. The N:P ratio should be a good indicator of nutrient limitation. The high N:P ratio results in P limitation, and vice versa and N is the limiting nutrient when this ratio is low. The optimum ratio is species specific, with the average being around 15 (Rhee & Gotham 1980; Stelzer & Lamberti 2001). However, the optimum N:P ratio for the tested strains of Stichococcus is not known. Elster & Komarek (2003) showed that the water of ultra- to oligotrophic Petrified Forest Creek has poor DIN, which results in a low N:P ratio. The lack of inorganic nitrogen could have a negative effect on periphyton growth in situ. Our laboratory experiments, based on a standardized method of AGP testing (Lukavský 1992), indeed directly suggested N limitation in the ultra- to oligotrophic samples. The amount of added N was enough to increase the N:P ratio in favour of higher microalgal production, due to greater P availability. In contrast, in the mesotrophic reach of Ornithologist Creek, where penguin excrement fertilizes the stream (mesotrophic water) through the addition of nitrates, periphyton growth was probably negatively influenced by a relative lack of phosphorus (low DRP concentration and high N:P ratio [Elster & Komarek 2003]). Also, further N enrichment of the mesotrophic water from Ornithologist Creek did not affect AGP, because the nitrogen content was already high and possibly even toxic.

The observed DRP concentrations are generally not limiting for microalgal growth (Grover 1989) and are comparable with oligotrophic and mesotrophic water types. In our experiments, the addition of P did not increase microalgal growth, since further additions decreased the low N:P ratio and N availability in the oligotrophic Petrified Forest Creek and the oligotrophic reach of Ornithologist Creek. The N:P ratio in the mesotrophic reach of Ornithologist Creek is one to two orders of magnitude higher than in the ultra- to oligotrophic waters. The addition of P could lead to increased microalgal productivity and decrease the extremely high N:P ratio caused by high N concentration from bird excrement. However, P enrichment did not result in the expected increase of microalgal production in our experiments, probably because the amount of added P (50 µg l⁻¹) was too low to substantially change the N:P ratio.

In contrast to the laboratory water trophic status evaluation, the higher primary production was observed in the oligotrophic Petrified Forest Creek and the oligotrophic reach of Ornithologist Creek (Elster & Komarek 2003). Several possible factors that could limit the periphyton growth in the mesotrophic reach of Ornithologist Creek were proposed: P limitation, toxic N concentration, presence of other toxic compounds, periphyton competition with guano-decomposing bacteria for nutrients, etc. (Elster & Komarek 2003). The possible P limitation, indicated by high N:P ratio, was not detected, but could not be ruled out either, by P enrichment in our experiments (see discussion above). High nitrogen concentrations were probably not toxic to the periphyton community in the field, since in our experiment the highest AGP₅ and AGP₂₅ values were observed in the mesotrophic reach of Ornithologist Creek.

Although data on nitrogen toxicity are not common, the effect depends on the chemical form of nitrogen and on the experimental strain. Generally, the growth inhibition occurs at nitrogen concentrations higher than 5–17 mg N l⁻¹ (Chu 1943). However, the range is much broader for different microalgae. For example, growth inhibition of Desmodesmus (Scencedesmus) quadricauda by nitrogen occurred at concentrations above ca. 300 mg N l⁻¹ (Přibil & Marvan 1970), while growth of Chlorella vulgaris was inhibited at concentrations above 31.5 mg l⁻¹ (Venkataraman 1969 and references therein). For the experimental Arctic Stichococcus strain, the inhibition should occur at concentrations above 25 mg DIN l⁻¹. No inhibition was observed in the temperate strain, indicating different nutrient requirements.

Any heavy metal growth inhibition could be detected by Micromethod of evaluation of water toxicity and AGP by a growth bioassay; TNV 757741 (Lukavský et al. 1995) and the presence of the heavy metals has been considered as one factor which could have decreased relative algal primary production measured in situ (Elster & Komarek 2003). Relatively high heavy metal concentrations were found in some inhabited Antarctic areas (Claridge et al. 1995; Sheppard et al. 1997). Selected Antarctic regions are also naturally enriched with Cd from the upwelling of marine sediment (Honda et al. 1987; Bargagli et al. 1996). Additionally, in the Antarctic, the negative effect of heavy metals could have been multiplied by other natural stress factors (Mayer et al. 1998). Mercury accumulation observed in penguin rookeries (Bargagli, Monaci et al. 1998) could have been one of the possible explanations for the lower productivity in that reach of Ornithologist Creek influenced by the penguin excrement (Elster & Komarek 2003). In our experiments, heavy metal inhibition was not identified in any of the oligotrophic and mesotrophic types of water samples, so heavy metal inhibition can be excluded as a cause of lower periphyton productivity in the mesotrophic reach.

The results of the AGP assay confirmed higher available nutrient content and excluded heavy metal inhibition in the reach of Ornithologist Creek influenced by penguin excrement, so other mechanisms for the depression of primary production in situ must be considered. Competition for nutrients between primary producers (microalgae and cyanobacteria) and guano-decomposing bacteria remains the most probable explanation for the decreased primary production in the mesotrophic part of Ornithologist Creek observed in situ (Elster & Komarek 2003).

The interactions between bacteria and primary producers (microalgae and cyanobacteria) are complex and influenced by environmental conditions, for example, nutrient input, nutrient stoichiometric imbalance or UV radiation, especially in oligotrophic environment (Cole 1982; Carrillo et al. 2002; Cotner & Biddanda 2002; Danger et al. 2007). Bacteria could inhibit phytoplankton growth due to competition for nutrients (Rhee 1972; Currie & Kalff 1984; Jansson 1993; Joint et al. 2002) or due to the production of inhibitory compounds such as antibiotics (Lemos et al. 1985), oxychlororaphine and 1-hydroxyphenazine (Dakhama et al. 1993) or spore germination inhibitors (Egan et al. 2001) that suppress algal growth. The presence of ecologically significant bacterial biomass in the mesotrophic reach of Ornithologist Creek, probably originating from the penguin rookery, is likely (Elster & Komarek 2003); these bacteria could out-compete the primary producers for available nutrients.

The strong correlation of the standardized values of AGP and the water nutrient load in oligotrophic and mesotrophic stream waters indicates that a standardized procedure for water trophic status determination, in this case AGP, can provide standardized reference points for a similar exploration of other shallow freshwater ecosystems across both polar regions (localities with similar and/or the same geological substrate type), which themselves form a natural gradient of climate change. Since the weathering rate is linearly correlated with average air temperature (Barnes et al. 2006; Convey et al. 2009), temperature increase will likely speed up the melting rate of glaciers and the weathering rate of geological bedrock. A study at eight Icelandic glaciated and non-glaciated catchments (Gislason et al. 2009) found an 8–30% increase in the mechanical weathering rate and a 4–14% increase in the chemical weathering rate for each Celsius degree of temperature increase. It seems probable that the weathering processes will affect the load and availability of nutrients to primary producers.

Using the method presented in this article, it should be possible to easily, accurately and comprehensively monitor the changing availability of mineral nutrients for cyanobacteria/microalgae growth, which is likely to occur with continued global warming. The field study of the chemical and biological water properties of hydro-terrestrial environments should combine laboratory measurements of the standardized AGP, and the results of both field and laboratory methods should be important indicators of the expected climate changes. Field studies include periphyton species composition and their spatial distribution, water mineral nutrients content, periphyton primary production and decomposition rates, zoobenthos composition and grazing rates. These results, especially those of repeated measurements at the same sites, should be stored in databases together with geographic information system data. Field observation data sets and standardized values of defined parameters, such as those proposed for the AGP values in this study, would become a valuable tool to assess changes or trends in spatial and/or temporal scales in polar streams. Long-term observations will provide data for models of the possible effects of climate change on polar hydro-terrestrial ecosystems.