For the Joint Manuela Action (Manuela proposal TASK 4) an experiment will be conducted rather than writing a review. One single field experiment will be executed in different regions.
The effect of one type of disturbance will be studied in different geographical areas. These selected areas are the Baltic Sea, The North Sea, the Atlantic and the Mediterranean. Since models predict a significant increase in rainfall as a consequence of global change, increased rainfall is selected as the effect to be studied.
Hypothesis: Sandy beach meiofauna from different geographic areas responds similarly to they same type and frequency of disturbance (i.e.increased input of freshwater as a result of changes in rainfall).
Restrictions in money and effort determined the nature of our research question; therefore it was decided to work with field experiments, in the intertidal, at sandy beaches.
The experimental design
With respect to spatial scale, 4 beaches, located in geographically separated areas, were designated as the core sites for the experiment. These are, in alphabetical order, Arina (microtidal beach, Mediterranean Sea, Krete), De Panne (macrotidal dissipative beach, North Sea, Belgium), Hel (microtidal beach, Baltic Sea, Poland), and Sines (macrotidal, dissipative beach, Atlantic coast, Portugal). The possibility of experimenting on additional sites in the UK and/or Wilhelmshaven was mentioned, but no commitments were made for these sites.
The basic sampling design would involve, on each site, 5 treatment and 5 control plots, organized in a randomised block design (5 blocks with 1 treatment and 1 control plot each). Plots would have a circular shape, and sampling within plots would follow a semi-randomized design. Distance between plots in a block should be large enough to avoid treatment effects on the control, and distance between blocks would typically be at least twice the distance between plots within a block.
Taking into account the different core sizes and numbers and the number of sampling events, and considering that ideally no more than 10 % of the plot area should be sampled in total, a first rough estimate suggests that the diameter of the sampling plots should be 2 metres. The surface of the rainfall plot should be larger (see figure below).
The experiment will be conducted after springtime i.e. in March 2007. The frequency of rainfall and timing of sampling is shown in the figure below. During the first week, rainfall will be applied daily. A last rainfall treatment will be conducted after T0 plus 3 weeks. As such, our treatment would mimic a one-week episode with intensive rainfall. The last treatment (single addition event) would serve to replicate the immediate response by the meiofaunal assemblage (see below).
Starting from T0, the experiment will be ended after 3 weeks. T-1 describes the initial situation and will be taken some time(s) (unidentified yet) in advance. A1 enables to quantify an immediate physiological response; A2 quantifies short term recovery; A3 quantifies the effect of repeated rainfall; A4 and A5 will measure the resilience of the communities. B samples are taken as a back up and archived.
On each sampling occasion, the following variables are sampled (1 core each from each experimental plot at each sampling occasion):
meiofauna – 5.5 cm diam. core , down to 15 cm depth, no vertical subdivisions
particle size analysis – 3.6 cm diam. core, down to 15 cm depth. A subsample for organic matter analysis will be taken from this core.
chl a - 3.6 cm diam. core, down to 2 cm depth.
porosity – 3.6 cm diam. core, down to 15 cm depth.
salinity – 5.5 cm diam. core, down to 15 cm depth, with vertical subdivision in 3 layers of 5 cm each
There was agreement that sediment permeability will also have to be determined, but not yet on whether this would actually be measured (and in that case, by whom) or approximated by calculations based on porosity and grain size data.
Sieve mesh size: lower limit will be 38 µm; no upper limit will be used. UGent will provide 4 (+ 1 extra) sieves.
The amount of fresh water to be added should resemble an intensive rainfall, applied within 10-15 minutes, i.e. 2.5 cm of water per m-2/day (this corresponds to x 10 British daily rain fall). This corresponds to 80 litre of water per plot; 400 litre every day (for the 5 replicate plots) and 3500 litre of water for the whole experiment (about 3000 litre for the first week).
The type of water to use: we aim for distilled water, however the different host institutes have to check whether it is feasible. The responsibles at the host institutes need to verify preparation and storage possibilities, and the costs of this amount of water. Alternatives can be tap water or rain water. There is some doubt in terms of standardization, this will be discussed further via email.
A watering can will be used to mimic rainfall (applied by several persons simultaneously).