Experimental design

Hypothesis test (i.e. Anova design) is used to study whether the response of a mesocosm unit differs from that of a control unit. Hypothesis tests are used for comparing averages and are characterised by having multiple replicates in control and treatment groups. The greater number of replicates, the greater is the power of the test for resolving differences.

Point estimate tests (i.e. Regression design) are used to evaluate regression relationships and, ideally estimate an exposure concentration which will not cause an adverse effect (NOEC or threshold concentration) or predict the intensity of an effect at a given exposure level. Regression design requires multiple treatments at various concentrations related to a response. The greater the number of treatment concentrations, the greater is the confidence in the fitted concentration-response line. As Point estimate tests assume a monotonic response of an effect parameter along a concentration gradient only direct effects can be evaluated. Even then indirect effects can mask the relationship.

Hybrid tests incorporate features of both hypothesis and point estimate tests by employing both multiple replicates and multiple doses. Fewer replicates will reduce the power to resolve significant effects and fewer dose levels will reduce the confidence in estimating the fit and the NOEC.

The majority of mesocosm experiments in the data base belong to the "Anova Design" or "Hybrid design" (6.2.1 & 6.4.1). We have evaluated the statistical power of the various designs by comparing the average reduction in abundance of zooplankton and macroinvertebrates at the lowest observed effect (significant) concentration (LOEC). Overall, the statistical power in the mesocosm studies was rather low. The average reduction in abundance of zooplankters exposed to insecticides at recorded LOECs was 75.4 % (± 21.3 %; SD) (see 6.2.1).

For macroinvertebrates the significant reduction was almost identical at 76.5 % (± 20.3 %; SD) (see 6.4.1). The low power is due to low number of replicates, low number of and/or large range in test concentrations.

Overall, the data suggest that in order to obtain a sufficient resolution and sensitivity the experimental design should be a hybrid design encompassing more than 4 test concentrations and at least two replicates at each concentration. As the size of experimental design usually is constrained by economic considerations with a maximum number of units of 15-18 they should be distributed between 5-6 test concentrations each with 2-3 replicates.

Therefore, in evaluating results from a mesocosm experiment one should take account of the experimental design, e.g. the results from a hybrid design with 5-6 test concentrations and 2-3 replicates each would produce the most reliable estimates of LOECs and NOECs.

Mesocosm design – size and depth

Mesocosms intend to mimic nature and ideally they should allow different groups of organisms to survive, behave and interact with other groups as in natural systems. Logistics and economy ultimately set limits to the maximal size that can be applied. If fish are to be included, systems need to be large, which invariably will impose patchiness and may introduce biases in the sampling procedure. Therefore, mesocosms of intermediate size are usually preferred.

The size of the mesocosm studies contained in the database varies widely. Systems with volumes from 0.003 m³ to 1,100 m³ and average depths ranging 0.1–5 m are included in the database, with the small systems primarily representing flow-through experiments. The influence of volume and depth on the sensitivity to pesticide exposure of different functional and taxonomic groups was tested using PLS analysis (Chapter 4).

The volume of mesocosm units had no influence on the toxicity of pesticides to either microalgae, zooplankton or macroinvertebrates (5.4.1, 5.5.1, 5.6.1), while the depth of the mesocosm significantly influenced the toxicity of insecticides to macroinvertebrates (5.4.1) with increasing effects (i.e. lower LOEC) at decreasing average depth of mesocosm.

Location and season of mesocosm tests

Length of growth season, solar insolation and temperature vary on a continuum of scales determined by geographical location and time of year. As each of these "external" variables affects populations of aquatic organisms (length of growth season: number of generations; insolation: algal growth; temperature: growth and metabolism) and the fate of pesticides (insolation & temperature: degradation) both the geographical location where mesocosm studies are carried out and time of year when carried out are expected to influence the expression of pesticide effects.

In the mesocosm studies contained in the data base neither temperature nor solar insolation are explicitly given for each sampling occasion. Therefore, we have used a sinusoidal function of the day no. to integrate these variables (e.g. day no. 183 attain the value 1 and day no. 1 and 365 attain the value 0).

All macroinvertebrate groups were most sensitive when the experiments were conducted at high latitudes.

Toxic effects are expected to occur at lower insecticide concentrations with increasing distance from Equator probably due to a slower turn-over of populations at high latitudes, i.e. fewer generations each year at lower temperatures (5.4.1). Consequently, recovery of macroinvertebrate populations after pesticide exposure takes longer time at northern latitudes.

For zooplankton effects of season and latitude of mesocosm was contradictory and no conclusion could be drawn.

Dosage of pesticides in mesocosms – single or multiple dosage

Pesticides enter the aquatic environment during field application as spray drift, in association with surface run-off during heavy rainfall and through subsurface run-off (e.g. drainage). The importance of the different routes of entry is rather specific to site, crop, method of application and physico-chemical characteristics of the pesticide. For these reasons mesocosm tests usually are tailored to answer specific questions and accordingly single dosage or multi-dosage of dissolved pesticides, or pesticides dosed in slurries have been applied. Such different application schemes make it difficult to compare the outcome of the various studies, as the application mode invariably will affect the concentration and fate of pesticides in the mesocosms, e.g. multiple dosing every week at a low concentration may result in a higher temporal-averaged concentration than a single dose containing an identical amount of pesticide.

At a given total dose effects of pesticides on macroinvertebrates will increase with interval between individual doses but decrease with number of doses. Therefore, a low but persistent pesticide concentration will have a lower effect on the macroinvertebrates than a high but temporary pesticide concentration (5.4.1). This may be due to the relativly long generation time of most macroinvertebrates. Hence, recovery will be hampered if pesticides are dosed at intervals close to the generation time.

For zooplankton the PLS models tested had the highest predictability (Q² = 0.736) when the only studies with a single addition of insecticide were included in the analysis (see 5.5). Inclusion of studies with multiple application of insecticides led to much lower goodness of fit and accordingly they were excluded in the analysis. Therefore, we cannot explicitly evaluate the influence of application mode on plankton.

Influence of sediment and macrophytes in mesocosms

Presence of sediment in a mesocosm should be a prerequisite for studying effects on macroinverte-brates. However, most zooplankters also rely on sediment for storage of resting eggs that constitute a "bank" for recolonisation.

Macrophytes are a natural component of shallow freshwater systems. They have an important structural role, providing habitat, shelter and food for a number of organisms, influencing the physical environment and, therefore, affect the biogeochemical fluxes near the sediments. Macrophytes may prevent sediment from erosion and resuspension, while promoting sediment deposition. In addition, macrophytes directly may influence the availability of pesticides by adsorption and uptake.

For macroinvertebrates the PLS analysis showed that the highest predictability (Q²(cum)) was obtained for a PLS model based on mesocosm experiments when both sediment and macrophytes were present in the test system (5.4). However, an almost similar high predictability was obtained for the PLS models for mesocosm experiments with sediment but without macrophytes in the test system. On the contrary, a much lower predictability was obtained when the PLS model was applied to all data for stagnant water including laboratory experiments without sediment.

Therefore, effects of pesticides on macroinvertebrates must be studied in mesocosms including sediment and preferentially also macrophytes in the test system. Omission of sediment in test systems may lead to erroneous results out of line with the majority of high-quality studies.

For zooplankton no consistent modifying effect of either sediment or macrophytes was found for the toxicity of pesticides (5.5).

Most sensitive groups – plankton - benthos

Aquatic organisms differ in their sensitivity to pesticides according to their taxonomy, generation time, functional role in the ecosystem and their habitat. Generally, non-target arthropods in aquatic habitats (crustaceans and insect larvae) are very sensitive to insecticides aimed to control insects in crops, while molluscs are considered less sensitive probably due to their ability to reduce exposure by shell closure.

Zooplankton: The PLS analysis showed that cladocerans are the most sensitive zooplankters to insecticides followed by copepods and rotifers (5.5.1). This was confirmed and detailed by regression analysis revealing that Cladocerans and Chaoborus are the most sensitive zooplankters followed by copepod nauplii and adult Copepoda (6.2).

The variation in sensitivity within each zooplankton group as demonstrated in mesocosm studies is considerable. Results from 3-4 detailed studies showed that LOEC varied 2 – 2.5 orders of magnitude within Cladocera (6.2). Hence, studies analysing Cladocera at the level of Order invariably will neglect effects on the species composition.

Macroinvertebrates: The PLS analysis showed no difference in sensitivity between predatory and non-predatory macroinvertebrates (5.4.1).

Detailed evaluation focussing on the sensitivity to insecticides of different taxonomic groups revealed that the insect order Tricoptera consistently was the most sensitive macroinverte-brate group, followed by Plecoptera /Hemiptera/Ephemeroptera/ Coleoptera/Amphipoda/Isopoda (6.4). Chironomidae as a very diverse group showed a rather large variation in sensitivity within a study (1-2 orders of magnitude). Hence, studies analysing effects on macroinvertebrates at the level of Order probably will neglect effects on the species composition.

Odonata and Gastropoda consistently were the groups with the lowest sensitivity to insecticides.

When comparing effects on zooplankton and macroinvertebrates the most sensitive organisms within macroinvertebrates generally will show lower LOEC than the most sensitive organisms within zooplankton (Chapter 7).

Therefore, mesocosm studies must include and focus on macroinvertebrates, as effects cannot be extrapolated from available single species tests (because macroinvertebrates are underrepresented in the single-species tests used for extrapolation of hazard concentrations). The macroinvertebrate community must include important and sensitive taxonomic groups such as Tricoptera, Ephmeroptera and Amphipoda.

Most sensitive effect parameter

Traditionally, mortality (and growth rate in algae) is the most widely used effect parameter in the regulatory procedure of pesticides because of ease of detection and obvious ecological significance. However, prior to death in an individual and reduction of a population sublethal effects will occur, which theoretically make sublethal effects excellent early warnings and sensitive effect parameters.

Abundance is by far the dominant effect parameter while functional effect parameters such as production and growth have seldom been measured and are therefore represented only at a limited scale (Chapter 4), which makes it difficult to compare the sensitivities.

The sublethal effect drift in stream macroinvertebrates generally appears to be a more sensitive endpoint than changes in abundance (6.4).

The endpoint emergence of adult insects generally is as sensitive as changes in abundance of larvae, however, sampling and interpretation can be difficult (6.4).

Duration of mesocosm experiments - recovery

Recovery of zooplankton populations following insecticide exposure relies on reproduction from surviving individuals, hatching of resting stages (eggs) or immigration. To be able to examine recovery of zooplankters, mesocosms therefore need to include sediment and, in addition, to be in operation for several weeks-months after pesticide dosing has stopped.

In macroinvertebrates recovery may take place by invasion from non-affected populations (e.g. by drift in streams, reproduction in insects) and reproduction by surviving individuals. In order to evaluate recovery, mesocosm studies need to be carried out in the field (to allow flying insects to lay eggs) and should at the minimum extend a full life cycle of the organisms studied after insecticide dosage.

Zooplankton: Less than 50 % of the mesocosm studies where zooplankton was followed, the post exposure period was too short and/or the doses of insecticides too high to observe complete recovery of zooplankton. For Cladocera the time elapsed for full recovery after the insecticide dosage varied between 10 and 120 days (6.2.2). In mesocosm experiments where Cladocerans had been reduced severely (i.e. > 95 %) it took more than 12-15 weeks for full recovery. At reductions below 80 % of the initial population size recovery was fast, less than 20 days. For copepods an almost identical relation between initial decrease and recovery was obtained. It should be noted that most recovery studies analysed the organisms at a "crude" taxonomic level (e.g. Cladocera). Therefore, recovery may take place by increase in "robust" species at the expense of sensitive species resulting in reduced species diversity and thus a decline of environmental quality.

Macroinvertebrates:

The majority of experiments in the database were terminated within 150 days. Taking the general life cycle length for macroinvertebrates into consideration (ranging from less than a month to several years), the experimental time frames in most mesocosm studies appear to be too short. Based on the few lengthy studies, Chironomids and Isopoda were the most important taxonomic groups in the "slight recovery group" whereas Chironomids and Ephemeropterans dominated the "moderate recovery group" (6.4.2). Both groups are considered as good colonisers with short life cycles and this probably explains why they show the most rapid recovery.