I think that a linear transect is correct as long as it'd be in good agreement with the aim of your sampling.

There are several concerns that come up to my mind:

First of them it's the purpose of your sampling: is it for monitoring a pollutant from an identified source?.

Secondly, I would consider the spatial and temporal scale of the hydrodynamic processes in the estuary, to determine both the direction of the transect (longitudinal or transverse) and the distance between sampling points.

On the other hand, I'd also think about the type of data you want to obtain and the statistical methods you'll use (from simple scatter plots to multivariate statistics).

I've found some publications that can illustrate this question:

- M Martino, A Turner, M Nimmo, G.E Millward, Resuspension, reactivity and recycling of trace metals in the Mersey Estuary, UK, Marine Chemistry, Volume 77, Issues 2–3, February 2002, Pages 171-186, ISSN 0304-4203

- Mills H, Kirby J, Holgate S, Plater A (2013) The Distribution of Contemporary Saltmarsh Foraminifera in a Macrotidal Estuary: an Assessment

of Their Viability for Sea-Level Studies. J Ecosys Ecograph 3: 131

-D. M. Sanger, A. F. Holland, G. I. Scott Tidal Creek and Salt Marsh Sediments in South Carolina Coastal Estuaries: I. Distribution of Trace Metals. Archives of Environmental Contamination and Toxicology. November 1999, Volume 37, Issue 4, pp 445-457

I hope this answer will be helpful to you

Regards

Gracias Raquel. Our goal is to monitor the environmental concentrations of enhanced naturaly ocurring radioactive materials (ENORM) discharged from a disabled phosphate plant and trace metals from local industries in the vicinity of the estuary. We'll be working in a wide mudflat in the low tide periods, sampling sediments (cores down to 60 cm), seaweeds (Zostera noltii) and invasive asian clams (Ruditapes philippinarum). The new thing here is that our fellow engineers are designing a 4wheel robot coupled to an aereal device and both will work on the recognition of sampling areas, local coordinates and sample collecting.

Gracias por la bibliografia que me recomiendas. Quizás volveremos a estar en contacto. Un saludo a la peña de ecologia de la UMA.

The question of the sampling point is always a debate. As raquel says, the aim of your sampling is important. But even more important is the number of samples that you can collect and measure. Sometimes this will be the major parameter, since your sampling scheme is not the same for 3 or 30 cores.

The best to start, if it's available, is to look at the grain-size variations. It's not really useful to sample areas with very similar grain-size, unless you expect a large role of proximal source of pollution.

Jose: Another consideration is if there is an environmental gradient that you might expect for the deposition of these constituents, you would then want to align your transect along the gradient (i.e., distance from the source, or perhaps elevation across the mud flat, etc.) that would give you information about whether such a gradient exists.

Olá Jose!

You can distribute your sediment sampling sites in a regular scheme ("blind" sampling strategy) or according to several environmental variables that you know before the sampling (e.g. grain-size, pollutants concentration).

Regarding to the quadrats or corers. This is a common but quite interesting question among ecologists. Both are acceptable in intertidal environments. However, some good reasons to use quadrats in mudflats surveys are mostly related to: (a) the majority of living organisms live in the upper sediment layer (remember that anoxic layer is close the bottom surface); (b) the possibility to compare your results with previous studies (in your study area or others with similar patterns/gradients).

Good luck!

Hi everybody. It's nice having all these contributions you made. I couldn´t find so far more information than 2 articles published by Carvalho 1995, 1997, in Sci Tot Environ, with relevant data on ENORM nuclides in water,suspended matter and bottom sediments. Any data we'll get from seaweeds and clams will be totaly new I guess. For the sediments we'll focuse on the upper layer (0-10 cm?) and a deeper theoretically preindustrial layer (45-55 cm?). Let´s see how it goes with the sampler robot.

Thank´s for your good advise and keep in touch.

A straifht transect will probably force you to go over areas with different microhabitats, submergion period and other variations. A more reasonable solution would be to make not one but 3 or 4 parallel transects. If you make them parallel to the coast who even obtain some info on the variation with topography. Even easier is to make a regular scheme or a "random" scheme which should avoid any biases from microvariations within the mudflat.

Hi Pedro. I could easily agree with the 3-4 parallel transects. I have been debating a lot about the sampling approach.What to do? A straight transect (perpendicular to the shore) every 3 or 4 sampling stations along the left bank was my first thought. Than, why not a regular scheme with the help of a grid set on the map of the region of interest. And now, what about 3-4 parallel transects every sampling station? Well, the working area extends roughly along 7km in a straight line along the shore and about 3km into the inner bay. That's a big space.But we'll have a robot and we'll be interacting with it and viewing what his videocameras are capturing on real time. So, the decision has to do with taking the best benefit of this newly designed tool. I know we will have a lot of a samples. But we must be wise enough when selecting the sampling spots and locations.

I thinnk the key is defining very well what are your questions. If you are interested in the effect of the distabce to the shore, then several transects perpendicular to the shore would be ideal. If you are interested in other variables that are not correlated with distance to the shore, than running transects paralel to the the shore will allow you to control for that effect, if you are interested in multiple variables, then the best choise is definitely either a regular scheme (you can adjust the grid size, i.e. distance between sampling stations, to get a reasonable sample size that is sufficient for statistics, but not to time/resource consuming) or a (pseudo-)random sampling scheme. I have used all these option in the past, the choice between them must be based on which questions you are trying to address.