Demographic variation and population viability in Gentianella campestris: effects of grassland management and environmental stochasticity

Citation
T. Lennartsson et Jgb. Oostermeijer, Demographic variation and population viability in Gentianella campestris: effects of grassland management and environmental stochasticity, J ECOLOGY, 89(3), 2001, pp. 451-463
Citations number
85
Categorie Soggetti
Environment/Ecology
Journal title
JOURNAL OF ECOLOGY
ISSN journal
00220477 → ACNP
Volume
89
Issue
3
Year of publication
2001
Pages
451 - 463
Database
ISI
SICI code
0022-0477(200106)89:3<451:DVAPVI>2.0.ZU;2-3
Abstract
1 Transition matrix models were used to evaluate the effects of environment al stochasticity and four different methods of grassland management on dyna mics and viability of a population of the biennial Gentianella campestris ( Gentianaceae) in species-rich grassland. Data were collected between 1990 a nd 1995. 2 Continuous summer grazing, the prevailing management strategy in Scandina vian grasslands, resulted in high recruitment of new plants, mainly because litter accumulation was prevented and gaps were created by trampling. Tram pling and repeated grazing, however, caused damage which reduced seed produ ction. Lambda for the average matrix was c. 0.77, and a stochastic matrix m odel yielded an extinction probability for the total population of c. 0.08 within 50 years. 3 Mowing in mid-July (used as a conservation tool) increased seed productio n, but litter accumulation following re-growth of the vegetation prevented establishment. Lambda and extinction risk were similar to continuous grazin g. 4 Mowing in October (another conservation tool) promoted recruitment becaus e of low litter accumulation, but the seed output decreased because plant g rowth was impaired by tall vegetation. Lambda was 0.64, while the extinctio n probability was very high (c. 0.98 within 50 years), 5 Mid-July mowing followed by autumn grazing (the historical management reg ime) yielded high values for both seed production and establishment of rose ttes. Lambda was 0.94 and the probability of extinction within 50 years was below detection level. 6 Log-linear analysis showed that the matrices differed significantly both between treatments and between years. The latter indicates environmental st ochasticity, here caused by summer drought that increased the extinction ri sk. Lambda may be slightly underestimated because drought occurred in one o ut of five summers during the study period, which is high compared with the natural frequency. 7 We conclude that traditional grassland management is more favourable for G. campestris than the methods that prevail in Scandinavia today. This indi cates a serious conservation problem, because grazing has replaced traditio nal management in many of the remaining semi-natural grasslands throughout Europe.