Impacts of reindeer grazing on soil properties on Finnmarksvidda , northern Norway

Numerous investigations have documented changes in vegetation due to reindeer grazing in Finnmark County, Northern Norway. However, rather few studies have focused on impacts o f reindeer grazing on soil properties. The aim o f this investigation was to identify possible changes in physical and chemical soil properties due to reindeer grazing. Furthermore, root distribution was detected. A t four different locations on Finnmarksvidda three sample sites each were selected subjectively according to lichen and plant cover: A) ample, B) reduced, and C) poor lichen and plant cover. It was supposed that differences i n lichen and plant cover were due to differences in reindeer grazing intensity. Results showed that the organic layer beneath ample lichen cover had an about 20% higher C E C and a 30—50% higher concentration o f plant available Ca and M g and total M g compared to those beneath reduced ones. A t sites with poor lichen and plant cover, an organic layer was mostly missing. The exposed mineral Eh-horizons at these sites had a significant (P<0.05) higher organic C content, higher C E C , concentrations o f total P , Ca and K , and plant available K , when compared to E-horizons beneath better lichen covers. Rooting depth and amounts o f plant available water i n the rooting zone were lower at sites with reduced and poor lichen cover. A relation was found between soil organic C and C E C for all soil samples, indicating that soil organic matter is an essential key factor for soil fertility at the investigate sites on Finnmarksvidda. Assuming that differences i n lichen and plant cover are related to differences i n grazing intensity, results indicate that overgrazing by reindeers can cause a significant degradation o f the organic layer, followed by significant losses o f essential plant nutrients, a reduction in plant available water and consequently soil fertility.


Introduction
Overgrazing is the world's furthermost reason for soil degradation (Oldeman et al, 1991).Also on Finnmarksvidda there is evidence that soil degradation takes place.Evans (1996) found that the organic horizons on less grazed sites were significantly thicker compared to more grazed ones at some sites, and that at some locations the organic horizon was "stripped off" the beneath mineral material.Trampling and grazing of lichens and plants by reindeer are today major factors controlling vegetation cover and composition all over the northern Fennoscandia (Våre et al., 1995;Stark et al., 2000;Kumpula, 2001).During the last three decades the investigation on the effects of RangHer, Special Issue No. 15,2004 reindeers on lichen dominated winter pastures has received special attention (e.g.Johansen & Karlsen, 2000).This is due to the importance of lichens as fodder for the survival of reindeers during winter.The condition of lichen ranges still form one of the most important economic bottlenecks for semidomesticated reindeer management in northern Europe {e.g.Kumpula, 2001).
The county of Finnmark, North Norway, manages by tradition, the greatest number of reindeer in Norway.From 1976From to 1988 the reindeer population size in Finnmark County increased from approximately 86 000 to 191 000 animals.Since then the number of animals decreased to about 144 000 animals in 1997 (Johansen et al., 1995;Ims, 1998) and less than 110 000 in 2001 (Haetta, 2002).Causes for these considerably decreases in reindeer population during the last decade are multifaceted, and combinations of historical, social, cultural, political, economical and natural reasons are discussed (e.g.Haugerud, 1999).
One commonly discussed cause is the unbalance between numbers of reindeer in relation to steady decreasing pasture resources.Evidence for reindeer grazing is widespread throughout Finnmark: from the almost ubiquitous presence of dung to the disappearance of lichens by trampling in summer grazings and by overgrazing in winter grazings (e.g.Johansen et al., 1995;Johansen & Karlsen, 1998); to the present of bare soil in many localities (Evans, 1996).Thus, there is widespread evidence for overgrazing by reindeers throughout Finnmark County.
Serious changes in soil structure and nutritional composition have been reported at locations with heavy trampling, as for example along reindeer fences (Evans, 1996;Olofsson et al., 2001).It has been general questioned, whether plants of the original vegetation still possess the ability to grow at these locations (Johansen & Karlsen, 1996).However, little is known about the impacts of reindeer grazing on soil properties of lichen dominated winter pastures.
The aim of this investigation was to document changes in physical and chemical soil properties due to reindeer winter grazing.

Selection of sample sites
The investigation was performed between the two sami villages of Kautokeino and Karasjok on the continental part of Finnmarksvida.The climate is characterized by a mean annual air temperature of -3.1 °C and an annual precipitation of 366 mm (Siccajavri; DNMI, 1999).Johansen & Karlsen (1996) characterized this area according to lichen cover as moderately grazed to overgrazed.However, remaining islands with apparently ample lichen cover are found solitarily.The following locations were chosen: 1) Suosijavrre I (69° 22.235'N, 24° 4.921'0); 2) Suosijavrre II (69° 21.849 N, 24° 23.080 E); 3) Lappoluobbal (69° 12.579' N, 23° 45.021' E); and 4) Bieddjucarro (69° 13.099 'N, 23° 46.713'E).At each location 3 sample sites (only two found at Lappoluobbal) were selected due to a subjective evaluation of the lichen and plant cover in accordance with the classification of Johansen & Karlsen (1996): A) ample, B) reduced, and C) poor lichen and vegetation cover.The sampling areas chosen were to be topographically and edaphically uniform, allowing us to assume that the vegetational differences between the sample sites are caused by grazing and not by environmental variation.Thus, to increase comparability between soil profiles all sample sites were established at locations with no or little inclination (0-2°) and a continuously morainic till as basic substrate.The morainic till were deposited by the glaciers of the Pleistocene glaciations and covers most of Finnmarksvidda (Olsen et al., 1996).Vegetation types chosen were lichen rich mountain birch forest (Suosijavrre I, Lappoluobbal and Bieddjucarro) and lichen rich dwarf shrub tundra (Suosijavrre II), which by area and usage represents important habitats for reindeer husbandry.

Soil description, sampling and analysis
At each sample site soil profiles with an approximately dimension of 100 cm x 100 cm x 80 cm were carefully excavated and thoroughly described by horizons according to FAO (1990) during the first weeks of September 1999 and 2000.Size, amount, orientation and distribution of roots were described in accordance to Hodgson (1976).The rooting zone was measured as the thickness of the O-horizon plus the depth of the mineral soil with root development.Due to the very stony composition of the morainic substrate, soil samples were taken from one open pit per sample site only.For chemical analysis soil samples were collected into polyethylene bags according to horizons and stored at approximately +5 °C until further treatments at the Chemical Analysis Laboratory of the Holt Research Centre.Prior to chemical and textural analysis, all samples were dried at 105 °C and passed through a 2 mm sieve to remove stones and large roots.Soil samples were analyzed for the following parameters: Particle-size distribution (Elonen, 1971), pore-size distribution (Richards, 1947;Richards, 1948), particle density (De Boodt et al., 1967), pHH2O (1:2,5 soil water suspension), organic carbon (Leco IG 212 carbon system), Kjeldahl-N (Bremner, 1960), base saturation (BS) and potential cation exchange capacity (CECpot) (Ogner et al., 1975;Ogner et al., 1977), plant available Ca, Mg, K and P (Egner et al., 1960).Analysis of filtered solutions were performed using a Perkin-Elmer Optima 3300 DV.Plant available water was calculated on basis of as the difference RangHer, Special Issue No. 15, 2004 between field capacity (lOOkPa) and permanent wilting point (l5OOkPa).

Statistical analysis
The number of soil samples in this investigation was not sufficient to check for normality in the data.However, in a similar study with fully comparable soil data, normality tests showed, that the investigated soil chemical parameters are normally distributed in general.Thus, analysis of variance was performed on all data using the GLM procedure of SAS (version 6.12).Chemical properties of the different soil horizons with different lichen cover were compared using the different locations as replicates.The individual means were compared using the Student-Newman-Keuls test.Significance was assigned to a test if P<O.O5.

Results
Soils were shallow podzols, stony with coarsetextured gritty matrix and a generally low clay fraction (Table 1).The soil texture was rather uniform, and all samples were loamy sands or sandy loams boardering to the loamy sand texture class.Soil densities were between 1.4 and 1.7 g/cm 3 , the soil porosity between 4O and 5O%, and air-filled porosity between 16 and 32% (data not shown) and there were rather little differences between sites.The pH in the soil profiles increased with increasing soil depth from about 4.O in the organic layer to approximately 6.O in the mineral C-horizon (Table 1).
The organic layer was generally thickest where lichen cover had the greatest thickness.Humus beneath ample lichen cover had a significant higher content of organic C and a higher CEC compared to humus beneath moderate lichen cover.There was a high correlation (r = O.99) between organic C and CEC (Fig. 1).Furthermore, humus beneath ample lichen cover had higher concentrations of plant available Ca and Mg and total Mg (Fig. 2a).Contents of Kjeldahl-N in the O-horizons varied between O.86-1.48%, and were thus about 1O-fold of those detected in the Eh-horizons (on weight basis).N concentrations in the B-horizons were in general below the detection limit of O.O5%.
No or only rather thin partial fragments of an earlier humus layer were found at sites with strongly reduced lichen cover.The accumulations of coarse mineral soil particles as coarse sand and gravel on the surface at these sites indicate that wind erosion has taken place to some extent.However, growth of pioneer lichens (e.g.Lecidea spp.) and mosses (e.g.Polytrichum spp.) had partly consolidated some of the exposed mineral soil.Characteristic for these sites were the occurrence of an Eh-horizon with a significant higher organic C content (3%) than Ehorizons (O.4%) beneath better lichen covers.Furthermore, Eh horizons contained higher concentrations of total P, Ca and K, plant available K (Figs. 2a, 2b), and a significant higher CEC.However, base saturations were generally below 5% (Table 1).
In B-and C-horizons no differences in physical and chemical soil properties between sites with different lichen cover were detected Growth of plant roots was concentrated in or directly below the O-layer with a mainly horizontal distribution.A more shallow root development and lower amounts of plant available water were found at sites beneath reduced and poor lichen cover (Fig. 3).Vascular plants like e.g.Empetrum nigrum or Vaccinium vitis-idaea growing at sites with poor lichen cover had a somehow stunted growth, and some chlorosis were observed.

Discussion
Due to no significant differences in chemical soil properties of C-and B-horizons, it was concluded that all investigated soils had developed from rather similar morainic till.The degree of podzolization, concerning the thickness of the E and B-horizons and the chemical properties, indicates further that soil genesis has taken place under rather similar conditions at all locations.If present differences in lichen and vegetation cover had existed for a longer period, more profound differences e.g. in pH or in the thickness of the characteristic podzolichorizons would be expected.The general similarity of all soil profiles indicate that differences in vegetation, and thus also the variability in soil characteristics of the uppermost soil horizons, are of rather recent character.Thus, based on the results about changes in lichens cover on Finnmarksvidda (Johansen & Karlsen, 1998) we assume that the most profound differences in soil properties between the investigated sites have occurred during the last 30 years.Levels of plant available Ca, Mg, K and P in the O-layer beneath ample lichen cover corresponded well with results reported from mountain birch forests in Kevo and Kilpisjarvi (Wielgolaski,2OO1).Differences in soil chemistry between sites with different lichen covers were limited to the O-and E-horizons, thus generally the uppermost 1O cm of soil profiles.The approximately 2O% lower organic C content, the about 27% reduction in CEC, and the between 3O to 5O% reduction in total and plant available Ca and Mg at sites with moderate lichen cover indicate, that considerable qualitative changes of the humus has taken place.A reduction in plant available Ca and Mg in the organic layer of about 5O% at sites with reduced lichen cover, is of a similar magnitude as reported for a North Finnish pine forest at heavily grazed sites (Vare et al., 1996).In contrast, no significant decrease of either total P or K could be detected in this investigation.
Despite the distinct differences between chemical soil properties of the organic layer beneath ample and reduced lichen cover, there were no differences in the mineral E-horizon.However, significant differences in chemical soil properties were found between the E-and Eh-horizons.The organic C contents in mineral E-horizons at sites with poor lichen cover were about 8-fold increased compared Rangifer, Special Issue No. 15, 2OO4 to sites with better lichen cover.Indications for the existence of a previous superficial O-layer at sites with poor lichen cover suggest, that the significant increase in soil carbon contents is a result of the incorporation of previous O-layer fractions with the underlying E-horizon due to grazing and trampling by reindeers.Differences in soil organic matter content are hold responsible for the detected differences in CEC and nutritious status between the Eh-and E-horizons.However, CEC of the mineral Eh-horizon was, on weight basis, less than 1O% compared to that of the O-layers.Furthermore, base saturation of Eh-horizons was in average 4% (Table 1), about 3 to 4 times less than in the O-layer.Poor root development and stunted vascular plant growth at heavy degraded sites indicate that losses of lichen cover and O-layer effect growing conditions negatively.
The decreasing thickness of the organic horizon with supposed increasing grazing pressure until its final loss, as also reported by Evans (1996), indicate considerable quantitative losses of soil organic matter.A high correlation (r = O.99) between the organic C content and CEC (Fig. 1) emphasize that the organic matter, in particular due to the rather low clay content (Table 1) of the morainic till, possesses a key function for soil fertility.
Consequently, any loss of organic matter would decrease soil fertility.There are no indications that the degradation of the superficial O-horizon has enhanced the nutritious status of mineral soils below, except for the E-horizons.Detected changes in C and nutritious status of Eh-horizons are significant, but too small to compensate losses from O-horizons.Therefore, it is assumed that a considerable amount of organic matter and nutrients were lost from sites with reduced and poor lichen cover.
Losses of organic matter and nutrients from sites with disturbed lichen cover could be partly due to erosion.Results from Fahnestock et al. (2000) indicate that the redistribution of litter by wind and snow during winter is an important mechanism of nutrient transfer across the arctic landscape.At sites with continuous vegetation cover redistribution of litter may be limited to above ground plant debris.However, once vegetation cover is disrupted, either by grazing, trampling or digging, also soil organic matter may get exposed to erosional forces, like e.g.deflation processes (Thannheiser, 1977).Evans (1996) reports that many of the descriptions of erosion in northern Scandinavia refers to the importance of wind erosion.Incidences of erosion due to reindeer grazing has been noted previously in Finnmark (Lyftingsmo, 1965;Vare etal, 1995;Johansen & Karlsen, 1998).
The declining amounts of actual plant available water in the rooting zone with reduced vegetation and humus cover indicates (Fig. 3), that reindeer grazing also has a significant impact on soil water budget.In general, a well developed lichen surface acts as an effective mulch and elevates the soil moisture status throughout the summer period by preventing evaporation (Larson & Kershaw, 1976;Kershaw, 1977;Kershaw, 1985).In subarctic welldrained terrain acute summer droughts are quite common which are assumed to be lethal to fine roots of trees and dwarf shrubs, that develop close to the soil surface (Crittenden, 2000).The observed reduction in root development (Fig. 3) may therefore to some extent be due to decreasing amounts of plant available water during the growing season.Thus, in addition to decreasing soil mineral nutrients, periodic drought may also be a significant factor suppressing vascular plant vigour.

Conclusions
The O-horizon posses a key role for soil fertility in the lichen dominated tundra-mountain birch forest on Finnmarksvidda.Over-grazing by reindeers can cause a significant degradation of the organic layer, followed by significant losses of essential plant nutrients, a reduction in plant available water and consequently soil fertility.Thus, over-grazing of winter lichen pastures may reduce vascular plant productivity and thus pasture quality and quantity.Further studies should focus on processes and extension of soils degradation due to reindeer grazing on Finnmarksvidda.

Table 1 .
Selected soil properties for representative pedons of the four study sites.O = organic, E = eluvial, B = illuvial horizon.