Any surveying around open water should be done with great care, and the use of boats is more dangerous. Unless you are out in the (shallow) river channel itself estimates of cover may be inaccurate because you are not looking down on the plants, so it may be useful to use a scoring system to avoid the spurious accuracy of percentage cover (see the ‘Quadrats’ section). Identification at a distance may also be difficult, and you may need to retrieve samples using a grab from the bank. Problems with cover estimates are described in the ‘Quadrats’ section, and solutions such as training of observers to decrease errors among observers are recommended. Grab sampling at depth will be extremely inaccurate and more technical solutions should be considered. Because rivers are rarely exactly linear, measures of length may vary depending on where they are made. Technically, you should measure along the midline (‘Thalweg’) of the river, but, even where it is possible to walk in the river, you would probably stir up sediment and so make subsequent surveying difficult. In practice, aim to measure along the bank, staying as close to the river as is possible and safe.