Inter-specific and inter-individual trait variability matter in surface 1 sediment reworking rates of intertidal benthic foraminifera

26 Although benthic foraminifera are an important component of meiofauna and contribute to 27 carbonate production and carbon/nitrogen cycles, their role in bioturbation processes remains 28 poorly known. Five dominant intertidal benthic foraminifera were recently classified into 29 functional bioturbator groups according to their sediment reworking mode and intensity. Our 30 study aimed at identifying potential drivers (i.e. size and/or travelled distance) of species31 specific surface sediment reworking rate. The travelled distance and surface sediment 32 reworking rate of Haynesina germanica, Cribroelphidium williamsoni, Ammonia tepida, 33 Quinqueloculina seminulum and Miliammina fusca were assessed through image analysis. Our 34 results show that the surface sediment reworking performed by these species is not size35 dependent, but dependent on their motility traits through interspecific differences in the 36 travelled distance. Smaller species (i.e. Quinqueloculina seminulum and Haynesina germanica) 37 contribute more to surface sediment reworking than larger ones (i.e. Ammonia tepida, 38 Cribroelphidium williamsoni and Miliammina fusca). These observations stress the critical role 39 of motion behaviour in surface sediment reworking processes by intertidal foraminifera. 40 Finally, we stress that the high inter-individual variability observed in conspecific motion 41 behaviour may be important to decipher the role of foraminifera in sediment bioturbation. 42 Noticeably, the species characterized by a strong inter-individual variability are also the species 43 that have the highest surface sediment reworking rates. This last observation may inform on the 44 species-specific phenotypic plasticity and therefore the potential for the functional role of these 45 species to be maintained in their natural environment. This is particularly relevant in an era of 46 global change where ecosystem balance is increasingly threatened by various stressors such as 47 heat-waves, ocean acidification and plastic pollution. 48 49 50

Molecular identification is therefore needed to discriminate morphospecies before the 129 assessment of their behavioural traits. In our sampling site, we find H. germanica S16, C. 130 williamsoni S1 and A. tepida T6 (Schweizer M., personal communication). Depending at the 131 abundance at the time of sampling, between 8 and 33 individuals per species with similar sizes 132 were used for the experiment (Table 1).  After the acclimation period, foraminifera were randomly placed on the sediment surface 165 (Fig. 2), under 250-ml of natural unfiltered and air saturated seawater (salinity = 33) following 166 10 min of vigorous air bubbling immediately before spreading foraminifera on the sediment 167 surface. In total 15 experiments were performed. Individual displacements in and on the sediment were recorded by time-lapse photography 174 ( Fig. 2; 1 image every 10 min for 24-hours) using a digital camera (Nikon V1 with a Nikkor 175 10-30mm lens). For each foraminifera, 144 images were combined in the image-analysis 176 software Fiji to extract (x,y) coordinates using the Manual tracking plugin (Schindelin et al. 177 into the sediment up to a depth where their paths were not visible and/or (ii) because some paths 180 crossed and consequently individual trajectories were lost. We therefore only kept individuals 181 that exhibited visible tracks throughout the whole 24-h experiment so that the information 182 related to their behavioural traits had the same statistical weight. In total we followed the 183 trajectories of 103 individuals. Note that at the end of each 24-h experiment, dissolved oxygen 184 saturation was consistently ca. 56% in the overlying seawater directly above the sediment-water  The distance travelled by each individual between two images (i.e. 10 min) was calculated 187 as: where (xt,yt) and (xt+10, yt+10) are the coordinates between two successive images taken at times   (Fig. 1a,b). Noticeably, the length/width ratio is close to 1 for A. tepida, 1.2 for H. germanica 206 and C. williamsoni and to 2 for Q. seminulum and M. fusca. The use of the surface area therefore 207 appears to be more accurate for the calculation. However, measurement of the cross section 208 involves the characterisation of the organism' height (Fig. 3). Due to their small size,

218
We therefore assume that the surface area corresponding to the larger part of the test i.e. reworking rate (SSRRit, mm 3 ind -1 10min -1 ) as follow: where Dt is the distance travelled (mm ind -1 d -1 ) by each individual between two images (i.e. 229 10min).

230
In our experiments, individuals of both species can moved on the surface (Fig 4A), at the 231 interface ( Fig 4B) or in the sediment ( Fig 4C).  Finally, the individual surface sediment reworking rate (SSRRi, mm 3 ind -1 day -1 ) were 244 estimated as follow:  individuals of Q. seminulum may range from 12 to 142 mm d -1 (Fig. 5).  typically is about 1.2 times larger than test width. As a consequence, the approach followed by  (Fig. 4).

328
Taken together, these suggest that considering both the surface area of the test as a proxy 329 of foraminiferal test size and the position of each individual in the sediment may allow a reliable 330 assessment of species-specific surface sediment reworking rate. 333 motion behaviour 334 Our results showed that the largest foraminiferal species (A. tepida) did not rework the 335 sediment more than the smallest species such as H. germanica and Q. seminulum. In contrast, 336 Q. seminulum can even rework up to 3 times more sediment than A. tepida (Fig. 6)   fusca prefer to be burrowed in the sediment that may likely induce the reworking of sediment 357 particles below the interface. As these species do not occupy the same microhabitat, they may 358 consequently exhibit difference in their SSRRi. Therefore, to consider the functional 359 classification of the five studied species in the estimation of their SSRRi, we include the position 360 of their test in the sediment.

361
Based on this study findings, it occurs that, although C. williamsoni is larger than H. 362 germanica and that both species displayed similar travelled distance, the latest rework the 363 surface-sediment more efficiently. This is consistent with previous studies which showed that  Noticeably, the species characterized by the highest inter-individual variability (i.e. Q. 397 seminulum, C. williamsoni and H. germanica) are also the species that contribute most to SSRRi 398 (see Fig. 6). This last observation is particularly relevant as it may inform on the species-399 specific phenotypic plasticity and therefore the sustainability of the functional role of these 400 species that live in a highly variable environment such as intertidal sedimentary ecosystems.