Agriculture, Ecosystems and Environment 115 (2006) 43–50 www.elsevier.com/locate/agee
Effects of changes in agricultural land-use on landscape structure and arable weed vegetation over the last 50 years Cornelia Baessler *, Stefan Klotz UFZ Centre for Environmental Research Leipzig-Halle Ltd., Department of Community Ecology, Theodor-Lieser Str. 4, D-06120 Halle, Germany Received 19 January 2005; received in revised form 5 December 2005; accepted 12 December 2005 Available online 24 January 2006
Abstract Agricultural dynamics and associated changes in the structure of habitat patches affect species composition and distribution in the landscape. Land-use, landscape changes and vegetation changes of weeds were analysed in a 4 km2 area in Central Germany (Saxony-Anhalt) from 1953 to 2000. This period includes the collectivisation (1952–1968), the agricultural industrialisation (1969–1989) and the privatisation of agricultural land following the political changes in East Germany in 1990. For the analyses, historic and current aerial photographs and vegetation data were used. Landscape indices and the average amount of mineral fertilizers were used as indicators for landscape structure and land-use intensity. Intensification of agriculture and the collectivisation in East Germany in the fifties and sixties led to a decline of the spatial heterogeneity of the landscape matrix (arable fields). The average number and cover of weed species, especially archaeophytes, decreased significantly since 1957. However, the total number of weed species increased. There was a remarkably high number of species with an average cover below 0.05%, called ‘‘chance’’ species in 2000. Out of 17 tested landscape indices only mean patch size and mean patch fractal dimension were significantly correlated with the average number of weed species. The average amount of the mineral fertilizer potash used as land-use intensity indicator was significantly negatively correlated with the total number of weed species. However, there was an increase in the number of farms after 1990 without changes in landscape structure and arable weed vegetation. The results suggest that structural variability of the landscape and habitat quality are the principal correlates of plant species diversity. # 2005 Elsevier B.V. All rights reserved. Keywords: Agricultural policies; Landscape change; Landscape indices; Land-use history; Vegetation change; Weeds
1. Introduction Agricultural land-use is dynamic and is related to changes in the structure of habitat patches, e.g. their spatial pattern, size or connectivity (Arx von et al., 2002; LaGro, 2001; Wagner et al., 2000). Agricultural intensification implies changes such as an increase in plot size of arable fields and the removal of linear elements. The resulting habitat isolation affects plant population dynamics and its basic processes at the landscape level, e.g. migration or colonisation. This is likely to play an increasingly important role for biodiversity patterns at the landscape level because many plant populations become isolated in otherwise * Corresponding author. Tel.: +49 345 5585 317; fax: +49 345 5585 329. E-mail address:
[email protected] (C. Baessler). 0167-8809/$ – see front matter # 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.agee.2005.12.007
unsuitable landscapes. This is why biodiversity studies need to be conducted also at the landscape level (Wiens et al., 1993). More than half of the territory of the E.U. is managed by farmers today (Vidalis and Lucas, 1999). The highest level of plant species diversity was reached in the 19th century, including many archaeophytes, known to be typical ‘weeds’ adapted to agricultural land-use (Ja¨ger, 1977). However, increasing agricultural intensification led to changes in landscape structure and thus in the composition and diversity of weed communities after 1950 (e.g. Medley et al., 1995). In East Germany, four periods of agricultural policies after the Second World War can be distinguished. During the first period from 1945 to 1952 a land reform was carried out (Eckart and Wollkopf, 1994). A state pool of the whole ground was created and small- and medium-sized farming
44
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50
units were developed during this time. During the second period (1952–1968) called ‘‘collectivisation’’, which was connected with the creation of the socialist state and of socialist public property, small farms were pooled to form large agricultural producers’ cooperatives (‘‘LPG’’). Only few companies with many employees handled a large area of arable fields and a high number of animals. In the third period (1968–1989) of ‘‘industrial agriculture’’ agricultural intensification increased, with progressive enlargement of farms associated with a separation of the farms for plant and animal production. However, 1990 marked another major turning point in agricultural policy, when the fourth period started with the privatisation of agricultural land following the political changes in East Germany. All agricultural producers’ cooperatives had been shut down by the end of 1991 (Eckart and Wollkopf, 1994) and since 1990, all German agricultural policies are subject to E.U. norms and regulations. So far numerous reports have dealt with the changes in the weed communities in Germany (e.g. Hilbig and Bachthaler, 1992; Otte, 1984). Only a few (e.g. Voigtla¨nder et al., 2001) however have included the period after the political change and thus the modified agricultural situation in East Germany since 1990. The objective of this study was to explore the influence of changing agricultural land-use and landscape structure on weed species richness in the last 50 years. We used data sets from the periods 1952–1968, 1969–1989 and after 1990 and focused on the following questions: What were the effects of changing agricultural land-use on the landscape structure, especially the arable fields? How did plant species numbers and species composition of agricultural habitats change? What are the main factors controlling species richness and composition in arable fields?
2. Material and methods 2.1. Study area The study area is located in the dry region of Central Germany near the village of Friedeburg (108340 E, 458120 N) and covers about 4 km2. It has subcontinental climatic conditions with a mean annual air temperature of about 9 8C (Veit et al., 1987) and an average annual precipitation below 500 mm. A plateau with nutrient rich loess deposits borders the steep slopes of the river Saale valley that forms the east border of the study area. The rivers Saale and Schlenze formed a wide floodplain with alluvial soil to the south of the study area. The geodiversity in the study area is coupled with high habitat and land-use diversity. The slopes are covered with woodland, meadows and pastures, and the areas of the floodplain and the plateau are under intense arable use.
2.2. Floristic data Floristic inventories of weed species of arable fields in the study area were available for the three periods from surveys performed in 1957 (Schubert and Mahn, 1959, 120 releve´s), 1979 (Westhus, 1980, 115 releve´s) and 2000 (220 releve´s). The current floristic composition of arable fields was documented by vegetation releve´s that were made from May to September 2000, just before harvesting of the different crop types. The releve´s (100 m2) were randomly placed in the total arable area. The sample plots were located at least 20 m from the field margins, because the agricultural conditions and thus the vegetation of this area often differ from the rest of the field (Elsen van, 1989). As in past inventories, the phytosociological method was followed according to Braun-Blanquet (1951) and Wilmanns (1989). Nomenclature of plant species follows Rothmaler (1994). 2.3. Land-use intensity and landscape structure As general indicators of land-use intensity the average amount of mineral fertilizers applied and farm size structure was used. Data for the three periods were taken for 1957, 1979 and 2000 from statistical yearbooks of the former German Democratic Republic (GDR) and the states of East Germany (Staatliche Zentralverwaltung fu¨r Statistik, 1960, 1980, 1987; Statistisches Landesamt Sachsen-Anhalt, 2000, 2001). Landscape structure was quantified by a set of landscape indices derived from land-use data. Data used for the three periods were extracted from aerial photos (black and white orthophotos) recorded in 1953 (1:22,000), 1969 (1:12,300) and 1997 (1:14,500). For the third period, we had to use the aerial photo recorded in 1969 because no aerial photo closer to 1979 was available. However, from the end of the 1960s until the beginning of the 1980s landscape structure did not change a lot (Schubert, 2001, personal communication). Minor changes of landscape structure between 1997 and 2000 were updated and ground truthing was performed by field mapping in 2000. To minimise possible interpretation errors the interpretation of all aerial photos was carried out by the same person. The land-use classification system included only seven types: woodland (including all woody habitats), meadows, dry and semi-dry grassland, arable fields, built-up areas, the river Saale and ‘‘others’’ (unclassified areas, total proportion <3%, Table 1). Based on the landscape elements (Forman, 1995; McGarigal and Marks, 1994) of the seven selected land-use types landscape indices were determined for each period. The landscape indices were calculated using FRAGSTATS, version 3.3 (McGarigal and Marks, 1994). In addition to the indices available in FRAGSTATS the index number of shape characterising points (NSCP) was used as a measure of shape and boundary complexity (Moser et al., 2002). The calculation of this index was carried out with an ArcView script developed by Moser et al. (2002).
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50 Table 1 Land-use intensity (amount of mineral fertilizers [kg ha1 a1]) used in the district of Halle, East Germany, from 1957 to 2000 (Staatliche Zentralverwaltung fu¨r Statistik, 1960, 1980, 1987; Statistisches Landesamt SachsenAnhalt, 2000, 2001), proportion of land-use types and structural traits of the arable fields in the study area Friedeburg, Central Germany from 1953 to 2000 1953/1957 Land-use intensity—fertilizer application Nitrogen (N) 34.9 Phosphorous (P2O5) 28.5 Potash (K2O) 71.1 Lime (CaO) 114.8 Proportion of land-use types Woodland (%) Meadows (%) Dry grassland (%) Arable field (%) Built-up area (%) River Saale (%) Others (%) Landscape structure of arable fields Number of fields Patch density (PD; x/100 ha) Mean size (MPS; ha) Smallest field (ha) Largest field (ha) Mean patch fractal dimension (MPFD) Number of shape characterising points (NSCP) Edge density (ED; m/ha)
1969/1979
2000
114.7 67.3 83.9 156
150 21.5 29.3 105.8
17.05 3.72 0.85 63.44 8.21 4.03 2.71
13.97 6.04 4.64 58.31 8.57 5.48 2.98
12.03 5.92 5.38 61.42 10.83 2.61 1.81
285 51.67 1.23 0.07 12.67 1.08
53 8.91 6.54 0.16 44.64 1.03
28 5.04 12.19 0.39 55 1.04
1354 320.08
318 87.26
45
The average of the accordant number received for each parameter and period was used. To calculate the average cover of each species over all releve´s of each period the cover code was replaced by the mean value of the respective cover class (Wilmanns, 1989). The differences of the average cover among periods was tested with a one-factorial variance analysis (ANOVA) and subsequent multiple comparison Scheffe´-test. The significance was evaluated using the Bonferroni correction by dividing the significance level ( p = 0.05) by the number of simultaneous tests (Legendre and Legendre, 1998). Only species with an average cover 0.02% in at least one period were included in the analysis (n = 130 species). The relationship between landscape indices and land-use intensity parameters and the total and average number of weed species was tested with Kendall’s tau rank correlation. Kendall’s tau represents a probability and will take values between 1 and +1, with a positive correlation indicating that the ranks of both variables increase together while a negative correlation indicates that as the rank of one variable increases the other one decreases. Kendall’s tau rank correlation was used instead of Spearman’s rank correlation because there is a direct interpretation of Kendall’s tau in terms of probabilities of observing concordant and discordant pairs (Conover, 1980).
318 90.52
3. Results 3.1. Changes in land-use and landscape structure
2.4. Statistical analysis A pairwise Wilcoxon-test was used to test for differences of average species numbers per releve´ among all periods. This test is optimal for the comparison of pairwise observations (Sachs, 1999). In contrast to the t-test, the Wilcoxon-test is independent from the data allocation, and it is very efficient for a small as well as for a large number of samples. Since there was an unequal number of releve´s per period, the total species number for each period was calculated by a resampling procedure. According to the lowest number of releve´s for all three periods, 115 releve´s were randomly chosen 1000 times from all releve´s of the respective period. Therefore, there were 1000 115 species lists for each period. The average of the total species numbers of these lists was used. To quantify weed species turnover, the number of: (1) common species, (2) species loss and (3) species gain between all periods was calculated. Data used for this analysis were available from the resampling procedure. The 1000 115 species lists of one period were compared with the corresponding species lists of the same resampling procedure of the other two periods irrespective of the number of common species, species loss and species gain.
The number of farms in East Germany heavily decreased between 1957 and 1960 from 825,124 to 20,317 and continued to decrease until 1979 (4816) with a concomitant increase in mean farm size (12–576–1231 ha). After the start of privatisation in 1990, the number of farms increased again (29,807 in 2000) accompanied by a decrease of farm size to 187 ha. The use of mineral fertilizers, i.e. phosphorous, potash and lime consistently increased from 1957 to 1979 and decreased again thereafter. In contrast, nitrogen application increased continuously from 1957 to 2000 (Table 1) with a pronounced increase between 1957 and 1979. Changes in land-use were confined to an increase in the area of meadows, dry grasslands and built-up areas and to a decrease in woodland. The proportion of agricultural area did not significantly change (Table 1). However, significant structural changes took place in arable fields, predominantly between 1957 and 1979 as shown in Fig. 1. There was a decrease in the number of patches, patch density, mean patch fractal dimension, the number of shape characterising points and edge density, and an increase in mean patch size, and in the size of the smallest and largest field (Table 1). This indicated a coarser landscape pattern, attributable to larger fields with a reduction of edges and linear elements associated with field delimitation.
46
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50
Fig. 1. Changes of landscape structure in the study area Friedeburg (Central Germany) from 1953 to 2000 illustrated by the patch boundaries.
The species with the highest cover differed among periods. In 1957, S. media was the most dominant weed species (3.50%), followed by C. album (1.55%) and P. aviculare (1.38%). In 1979, G. aparine reached the highest cover (4.34%), followed by C. album (2.26%) and S. media (1.76%). In 2000, the lowest total cover was reached. The species with the highest cover reached only 0.70% (C. album) followed by F. convolvulus (0.44%) and P. aviculare (0.42%). In 2000, there was also a clear increase in the number of species with very low average cover. Three-quarter of the species had an average cover below 0.05%. Most of these were either typical agricultural weed species like Hyoscyamus niger, Knautia arvensis or Sherardia arvensis, which became really rare or species typical for other adjacent habitats, like Lepidium ruderale, Lysimachia nummularia or Tanacetum vulgare. The cover of 23 species, out of the 130 tested, changed significantly over the period (Table 2). While the cover of 16
3.2. Changes in weed species numbers and composition There was a highly significant decrease ( p < 0.001, Wilcoxon-test) of the average number of weed species per releve´ from 20 species in 1957 to 14 species in 1979 (decrease of 30%). Since 1979, there was no significant change (15 species in 2000). There is also a highly significant difference between the average number in 1957 and 2000 ( p < 0.001). The total number of weed species (calculated after resampling) increased from 128 species in 1957 to 187 in 2000 (121 species in 1979). Although the weed flora changed among periods, all crop types were dominated by the same few species: Stellaria media, Chenopodium album, Fallopia convolvulus, Polygonum aviculare, Galium aparine, Amaranthus retroflexus and Cirsium arvense. Weed cover decreased substantially during the whole time (average values per period: 0.21–0.15–0.04%).
Table 2 Weed species with significant changes in average cover between the periods 1957, 1979 and 2000 (after ANOVA and Scheffe´-test with Bonferroni correction) Plant species
Amaranthus retroflexus Arenaria serpyllifolia Capsella bursa-pastoris Consolida regalis Descurainia sophia Euphorbia esula Euphorbia exigua Euphorbia peplus Galium aparine Lamium amplexicaule Lithospermum arvense Mentha arvensis Papaver rhoeas Plantago intermedia Senecio vulgaris Silene noctiflora Sinapis arvensis Sonchus oleraceus Stellaria media Thlaspi arvense Veronica agrestis Veronica hederifolia Veronica polita
Scheffe´-test ( p)
Average cover 1957
1979
2000
1957–1979
1957–2000
1979–2000
0.01 0.49 0.42 1.09 0.39 0.02 0.47 0.19 0.21 0.35 0.21 1.17 0.60 0.00 0.15 0.33 0.63 0.39 3.43 0.25 0.40 0.99 0.00
0.87 0.00 0.08 0.03 0.03 0.00 0.03 0.07 4.34 0.13 0.01 0.05 0.03 0.05 0.01 0.21 0.17 0.22 1.76 0.02 0.00 0.18 0.18
1.90 0.00 0.05 0.04 0.16 0.00 0.19 0.02 0.39 0.02 0.01 0.08 0.34 0.00 0.02 0.02 0.00 0.02 0.55 0.01 0.00 0.00 0.35
n.s. n.s.
n.s. n.s.
**
**
**
**
**
n.s. n.s. n.s. n.s. n.s.
n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
n.s., not significant at p 0.00038; all ANOVAs were significant. * p 0.00038. ** p 0.00007.
n.s. **
n.s. **
n.s.
**
**
**
**
**
*
n.s.
n.s. n.s.
**
**
n.s. n.s. n.s. n.s.
**
**
**
**
**
**
**
n.s.
**
** ** **
**
n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50
47
Table 3 Average number of common, gained and lost species between the sequent time periods with resampling (1000 permutations) Time period
Average number of common species
Average number of species gain
Average number of species loss
1957/1979 1979/2000
52.6 57.6
28.4 48.4
36.7 23.3
1957/2000
50.0
56.0
39.3
species declined and of two species increased significantly, only five species showed fluctuating values from 1957 to 2000. Typical weed species showing a highly significant decrease were, e.g. Consolida regalis, Lithospermum arvense, Mentha arvensis, Silene noctiflora, S. media and Thlaspi arvense. In contrast, A. retroflexus and Veronica polita showed a significant increase in cover. The weed species turnover was mirrored in the changes of weed species numbers (Table 3). There was nearly the same number of common species among all periods. Nevertheless, the lowest number of common species (50) was found when comparing the weed species of 1957 and 2000. Many species were lost after 1957. However, there is a clear increase in the number of species gain from the first (28.4 species) to the second period (48.4 species). 3.3. Relating weed species richness to land-use and landscape structure The correlation between weed species numbers and landscape indices as well as land-use parameters is shown in Table 4. There was no significant correlation between any landscape index and total number of weed species. Six landscape indices were positively correlated to the average
number of weed species (Kendall’s tau: 1.00; percent of landscape, mean shape index, area-weighted mean shape index, mean patch fractal dimension, area-weighted mean patch fractal dimension and sum of the number of shape characterising points). Furthermore, the land-use parameter lime was negatively correlated to the average number of weed species and phosphorous and potash were negatively correlated to the total number of weed species (Kendall’s tau: 1.00). 4. Discussion The main results of the study were that: (1) the spatial heterogeneity of the landscape matrix of arable fields declined significantly through time, (2) the average weed species number per releve´ and the average weed species cover decreased significantly; especially typical weed species (archaeophytes) decreased; whereas the total number of weed species increased and (3) the main factors influencing species richness on the arable fields are the complexity of the landscape matrix and thus patch size in conjunction with the number of field boundaries, and land-use intensity, especially the application of mineral fertilizer.
Table 4 Kendall’s tau rank correlations of landscape indices and land-use parameters vs. weed species numbers; data from 1957, 1979 and 2000 Index/parameter CA %LAND NUMP MPS PSCOV PSSD TE ED MPI MSI AWMSI MPFD AWMPFD PD SUMNSCP (whole area) NSCP (arable fields) AWNSCP Nitrogen Phosphorous Potash Lime
Total class area Percent of landscape Number of patches Mean patch size Patch size coefficient of variation Patch size standard deviation Total edge Edge density Mean proximity index Mean shape index Area-weighted mean shape index Mean patch fractal dimension Area-weighted mean patch fractal dimension Patch density Sum of the number of shape characterising points Number of shape characterising points Area-weighted mean of NSCP
Total species number
Average species number
0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.33 0.00 0.33
0.33 1.00 0.33 0.33 0.33 0.33 0.33 0.33 0.33 1.00 1.00 1.00 1.00 0.33 1.00 0.82 0.33
0.33 1.00 1.00 0.33
0.33 0.33 0.33 1.00
48
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50
4.1. Changes in landscape structure The intensification of agricultural land-use led to largescale changes in the landscape structure of the study area after 1950. Similar changes are generally found in Europe (Ihse, 1995; Voigtla¨nder et al., 2001) and in North America (Medley et al., 1995). A fine-grained land-use mosaic is replaced by large homogenous area as small fields are joined, linear elements are reduced and thus, boundary vegetation is removed. However, despite a drastic decline in landscape complexity and spatial heterogeneity, the share of land-use types within the whole landscape did not drastically change over the period investigated (Table 1). The most radical changes in the structure of the arable fields in the study area occurred during the period 1953– 1969 (Table 1; Fig. 1). Many small fields were replaced by few large fields in conjunction with the reduction of linear elements. Voigtla¨nder et al. (2001) reported similar developments for different sites in East Germany between 1950 and 1970. These changes were directly connected with the collectivisation and the increasing intensification of agriculture. Only marginal changes in the structure of the fields took place in the following two periods up to now despite considerable changes in farm size structure, i.e. increased number of farms and decrease in farm size caused by the changes in agricultural policies after 1990. 4.2. Changes in weed species diversity The average number of weed species was still relatively high in 1957. Although a merging of fields was initiated at this time, a differentiation of weed communities according to abiotic site conditions was still given (Schubert, 2001, personal communication). Furthermore, a time lag can be expected between the structural and the biological dynamics in the field so that changes of spatial structures and of species distribution patterns are not closely synchronised (Lindborg and Eriksson, 2004). In addition, tillage was still less intense in the 1950s so that crop residues were left on the fields. Crop residues help to trap wind-borne seeds or favour weed establishment (Tuesca et al., 2001), resulting in a higher density of wind-dispersed species on fields with reduced tillage (Derksen et al., 1994). As a result of agricultural intensification, the average number of weed species decreased significantly up to 1979. There was only a slight fluctuation of species numbers later on. Similar developments were observed in other areas of Germany and Europe (Andreasen et al., 1996; Hilbig and Bachthaler, 1992; Otte, 1984; Voigtla¨nder et al., 2001). Borowiec (1988) confirmed these developments for Poland by comparing small individual farms with management practises resembling those of the fifties in Central Europe, to intensively used farms with large structures. A similar trend could be observed for the average cover of weed species. Many archaeophytes, frequently surveyed in the past and known to be typical weeds adapted to
agricultural land-use, decreased significantly in their cover or vanished entirely from the fields since 1957 (e.g. Adonis aestivalis, Anthemis arvensis, Centaurea cyanus, C. regalis, L. arvense, M. arvensis, Scleranthus annuus, S. arvensis, S. noctiflora). These and other weed species were not able to keep pace with the rapid agricultural intensification of the last century (Voigtla¨nder et al., 2001). Some of them retreated into remnants of heterogeneous, semi-natural habitats or even into urban areas and abandoned their original close connection to agricultural areas (Deutschewitz et al., 2003). In fact, a number of the above-mentioned weed species are nowadays restricted to field margins in the study area and are missing in the centre of fields. Thus, marked differences in species diversity have developed between the core area and the edge of fields (Elsen van, 1989). This is probably related to the more intensive agricultural use in the core area (Wagner et al., 2000). After 1990, no come-back of typical weeds from the field margins into the fields could be observed, despite a partial reduction of fertilizer application and changes in farm size structure, which however, did not result in changes in landscape structure. In 2000, a remarkably high number of species with an average dominance below 0.05%, called ‘‘chance’’ species, were found. Many of these are typical weed species like S. arvensis, K. arvensis or H. niger, which became extraordinarily rare. In addition, species typical for other, adjacent habitats, like L. ruderale, T. vulgare or L. nummularia were observed. Many species of the last group are ruderal species, which are opportunistic and use gaps between otherwise very dense crop plants. The increase in the total number of weed species in 2000 is due to a relatively large number of these ruderal species with low abundance. 4.3. The role of land-use intensity and landscape structure on weed species richness Habitat quality is the principal determinant of vascular plant species diversity (Griffiths et al., 2000). In an agricultural landscape with a small-scale mosaic structure, habitat quality is mainly determined by the anthropogenic factors land-use type and intensity, while the dominant natural environmental factor is soil topography, which integrates both the nutrient supply and hydrological site conditions (Waldhardt et al., 2004). Thus, an enlargement of the fields and the interferences in water balance and soil structure have led to a levelling, uncoupling the connection between land-use and soil type (Ihse, 1995). The overlap of abiotic site conditions with land-use has an important influence on species diversity. In the study area, habitat specialists like C. regalis, L. arvense and S. arvensis are receding and are being increasingly replaced by generalists like S. media, C. album and G. aparine. The results suggest that the intensification of land-use, especially the application of mineral fertilizer, affects habitat quality and thus leads to the homogenisation of the flora in fields.
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50
Another negative influence on weed species diversity in the fields is the higher density of crop plants obtained from intensification methods (Weiner et al., 2001). In contrast, the variety of crops represents a diversity component among fields (Wagner et al., 2000). Nevertheless, this effect was almost completely deleted by the levelling of site conditions and particularly by the reduction of crop rotations to only a few cultivated plants during the 1970s. However, the less intense agricultural land-use together with the diverse crop rotation and the diversified cultivation methods practised up to the 1950s promoted a species rich weed flora until the end of the 1950s. Floristic diversity is furthermore mainly determined by the structural variability of the landscape, which is influenced by human interference (Forman and Godron, 1986; Wagner et al., 2000). Generally, increased variability of the spatial structure due to anthropogenic use has a positive effect on floristic diversity (Miller et al., 1997). However, there is a negative influence if anthropogenic landuse leads to a coarser landscape pattern with large fields dominating the landscape (Deutschewitz et al., 2003). The reduction of linear elements associated with field merging reduces the possibility to retreat for many weed species (Hovd and Skogen, 2005). A negative effect of land-use intensification on biodiversity in agricultural landscapes was identified by several studies (e.g. Luoto, 2000; Zechmeister and Moser, 2001) and was interpreted as being due primarily to a decreasing diversity in land-use patterns and semi-natural and natural habitats and the increasing size of arable field patches. In our study, the changes in weed species richness were positive correlated with the landscape indices %LAND, MSI, AWMSI, MPFD, AWMPFD and SUMNSCP (Table 4). These results reveal that decrease of landscape complexity in conjunction with enlargement of arable fields and reduction of linear elements lead to a decrease in the average number of weed species. In conclusion, weed species richness is negatively influenced by decreased structural diversity and landscape heterogeneity. These small-scale effects may affect biodiversity patterns on larger scales. Comparing grid cells (120 km2) Deutschewitz et al. (2003) have shown that plant species richness, including archaeophytes, was higher in grid cells with small-scale land-use than in cells dominated by large agricultural patches. To conclude, the less intense agricultural land-use during the first period of agricultural policies in East Germany (1945–1952) promoted a high spatial heterogeneity of the landscape matrix and a species rich weed flora. There was still high species diversity at the beginning of the second phase (1952–1968). However, the enlargement of arable fields together with reduction of linear elements during this phase and the increase in agricultural intensification in the third phase (1968–1989) led to a decrease in spatial heterogeneity and a decrease in weed species diversity in the fields. After 1990, there were no significant changes in landscape structure. Although the total number of weed
49
species increased in 2000, the average weed species number per releve´ did not change. Thus, the return of typical weed species into the fields depends not only on the partly decreased land-use intensity, but also on higher structural diversity, e.g. linear elements in the area of arable fields.
Acknowledgements The research was supported by the Energy, Environment and Sustainable development Program (FP5) of the European Commission (contract number EVK2-CT-200000082). We are grateful to R. Schubert, E.-G. Mahn and W. Westhus for their releve´s of the years 1957 and 1979, and D. Moser for making the script of the landscape index NSCP available. We thank A. Lausch for helps in GIS and I. Ku¨hn for discussion on statistical analyses. In addition, we are grateful to W. Durka for his contributions and S. Gwillym for improving linguistics.
References Andreasen, C., Stryhn, H., Streibig, J.C., 1996. Decline of the flora in Danish arable fields. J. Appl. Ecol. 33, 619–626. Arx von, G., Bosshard, A., Dietz, H., 2002. Land-use intensity and border structures as determinants of vegetation diversity in an agricultural area. Bull. Geobot. Inst. ETH 68, 3–15. Borowiec, S., 1988. Erfassung und Bewertung anthropogener Vera¨nderungen in segetalen Gemeinschaften Nordwest-Polens. Math. -Nat. Reihe, Wiss. Z. Univ. Halle 37, 127–136. Braun-Blanquet, J., 1951. Pflanzensoziologie. Grundzu¨ge der Vegetationskunde. Springer, Wien. Conover, W.J., 1980. Practical Nonparametric Statistics. New York. Derksen, D.A., Thomas, A.G., Lafond, G.F., Loeppky, H.A., Swanton, C.L., 1994. Impact of agronomic practices on weed communities: fallow within tillage systems. Weed Sci. 41, 409–417. Deutschewitz, K., Lausch, A., Ku¨hn, I., Klotz, S., 2003. Native and alien plant species richness in relation to spatial heterogeneity on a regional scale in Germany. Global Ecol. Biogeogr. 12, 299–311. Eckart, K., Wollkopf, H.F., 1994. Landwirtschaft in Deutschland. Vera¨nderungen der regionalen Agrarstruktur in Deutschland zwischen 1960 und 1992. Beitra¨ge zur Regionalen Geographie. Institut fu¨r La¨nderkunde, p. 36. Elsen van, T., 1989. Ackerwildkraut-Gesellschaften herbizidfreier Ackerra¨nder und des herbizidbehandelten Bestandesinneren im Vergleich. Tuexenia 9, 75–105. Forman, R.T.T., 1995. Some general principles of landscape and regional ecology. Landscape Ecol. 10, 133–142. Forman, R.T.T., Godron, M., 1986. Landscape Ecology. John Wiley & Sons, USA. Griffiths, G.H., Lee, J., Eversham, B.C., 2000. Landscape pattern and species richness; regional scale analysis from remote sensing. Int. J. Remote Sens. 21, 2685–2704. Hilbig, W., Bachthaler, G., 1992. Wirtschaftsbedingte Vera¨nderungen der Segetalvegetation in Deutschland im Zeitraum von 1950–1990. Angewandte Botanik 66, 192–200. Hovd, H., Skogen, A., 2005. Plant species in arable field margins and road verges of central Norway. Agricult. Ecosyst. Environ. 110, 257– 265. Ihse, M., 1995. Swedish agricultural landscapes—patterns and changes during last 50 years, studied by aerial photos. Landscape Urban Plan. 31, 21–37.
50
C. Baessler, S. Klotz / Agriculture, Ecosystems and Environment 115 (2006) 43–50
Ja¨ger, E.J., 1977. Vera¨nderungen des Artenbestandes von Floren unter dem Einfluß des Menschen. Biologische Rundschau 15, 287–300. LaGro, J.A., 2001. Landscape Ecology. Nature. Encyclopedia of Life Sciences. Macmillan Publishers Ltd., England. Legendre, P., Legendre, L., 1998. Numerical Ecology. Developments in Environmental Modelling. Elsevier, p. 20. Lindborg, R., Eriksson, O., 2004. Historical landscape connectivity affects present plant species diversity. Ecology 85, 1840–1845. Luoto, M., 2000. Modelling of rare plant species richness by landscape variables in an agriculture area in Finland. Plant Ecol. 149, 157– 168. McGarigal, K., Marks, B., 1994. Fragstats—Spatial Pattern Analysis Program for Quantifying Landscape Structure. Forest Science Department, Oregon State University, Corvallis. Medley, K.E., Okey, B.W., Barrett, G.W., Lucas, M.F., Renwick, W.H., 1995. Landscape change with agricultural intensification in a rural watershed, Southwestern Ohio, USA. Landscape Ecol. 10, 161–176. Miller, J.N., Brooks, R.P., Croonquist, M.J., 1997. Effects of landscape patterns on biotic communities. Landscape Ecol. 12, 137–153. Moser, D., Zechmeister, H.G., Plutzar, C., Sauberer, N., Wrbka, T., Grabherr, G., 2002. Landscape patch shape complexity as an effective measure for plant species richness in rural landscapes. Landscape Ecol. 17, 657–669. ¨ nderungen in Ackerwildkraut-Gesellschaften als Folge Otte, A., 1984. A sich wandelnder Feldbaumethoden in den letzten 3 Jahrzehnten. Dissertationes Botanicae. J. Cramer, Berlin, Stuttgart, p. 78. Rothmaler, W., 1994. Exkursionsflora von Deutschland, Gefa¨ßpflanzen: Kritischer Band. Gustav Fischer Verlag Jena. Sachs, L., 1999. Angewandte Statistik. Anwendung statistischer Methoden. Springer. Schubert, R., Mahn, E.G., 1959. Vegetationskundliche Untersuchungen in der mitteldeutschen Ackerlandschaft, I. Die Pflanzengesellschaften der Gemarkung Friedeburg (Saale) Wiss. Z. Univ. Halle, Math. -Nat. Reihe. VIII, 965–1012. Staatliche Zentralverwaltung fu¨r Statistik. Statistisches Jahrbuch der DDR 1960/1980/1987. 5. Jg./32. Jg./35. Jg. Deutscher Zentralverlag, Berlin.
Statistisches Bundesamt, 2000. Du¨ngemittelversorgung. Produzierendes Gewerbe, Fachserie, p. 4. Statistisches Landesamt Sachsen-Anhalt. Statistisches Jahrbuch des Landes Sachsen-Anhalt 2000/2001. Tuesca, D., Puricelli, E., Papa, J.C., 2001. A long-term study of weed flora shifts in different tillage systems. Weed Res. 41, 369–382. Veit, U., Petzold, B., Piehl, H.-D., 1987. Klimadaten der DDR—Ein Handbuch fu¨r die Praxis. Klimatologische Normalwerte 1951/80. Reihe B. Meterologischer Dienst der DDR, Potsdam. Vidalis, C., Lucas, S., 1999. Europa¨ische Landschaften: Mehr als die Ha¨lfte der Fla¨che wird landwirtschaftlich genutzt. ‘‘Statistik kurzgefaßt’’. Landwirtschaft, Eurostat. Voigtla¨nder, U., Scheller, W., Martin, C., 2001. Ursachen fu¨r die Unterschiede im biologischen Inventar der Agrarlandschaft in Ostund Westdeutschland. Angewandte Landschaftso¨kologie. Bundesamt fu¨r Naturschutz, Bonn-Bad Godesberg, p. 40. Wagner, H.H., Wildi, O., Ewald, K.C., 2000. Additive partitioning of plant species diversity in an agricultural mosaic landscape. Landscape Ecol. 15, 219–227. Waldhardt, R., Simmering, D., Otte, A., 2004. Estimation and prediction of plant species richness in a mosaic landscape. Landscape Ecol. 19, 211– 226. Weiner, J., Griepentrog, H.W., Kristensen, L., 2001. Suppression of weeds by spring wheat Triticum aestivum increases with crop density and spatial uniformity. J. Appl. Ecol. 38, 784–790. Westhus, W., 1980. Die Pflanzengesellschaften der Umgebung von Friedeburg (Kr. Hettstedt) und Wanzleben wa¨hrend des Zeitraumes 1978/79 und ihr Vergleich mit Untersuchungsergebnissen von 1958/59 bzw. 1961/62. -Diplomarbeit. Martin-Luther-Universita¨t Halle-Wittenberg. Wiens, J.A., Stenseth, N.C., van Horne, B., Ims, R.A., 1993. Ecological mechanisms and landscape ecology. OIKOS 66, 369–380. ¨ kologische Pflanzensoziologie. Quelle & Meyer, Wilmanns, O., 1989. O Heidelberg. Zechmeister, H.G., Moser, D., 2001. The influence of agricultural land-use intensity on bryophyte species richness. Biodivers. Conserv. 10, 1609– 1625.