In the 1940s a British scientist by name of Beauchamp located a fixed sampling station for physicochemical parameters 5 km north-east of Nkhata Bay in the northern offshore waters of Lake Malawi. The lake is about 400 metres deep at that Beauchamp's sampling station, but no details were given on why that location was considered representative of the whole lake. I'm intending to do my sampling at that same location, as several years have lapsed after Beauchamp did his work and temporal changes may have occurred since then, but in absence of clear reasons why that one location was considered representative I tend to hesitate.
You have reason to hesitate Elias, because is very hard make a sample a representative one. If it is unavoidable the necessity to stablish a fix point it does not mean you can verify previously and after the stablishment the stability of the lake, you can take random samples and check if there is homogeinity in those, and if there is no special issues like zones near to the shore (where the contact with the soil and animal can make chmical differences) or perhaps zones where certain animal prefers to stay and its activity changes the profile. If you find a high level of stability perhaps your study will be representative, if it was not best you can do is continue checking your fix point as reference and random samples to contrast and have a more critical and deep knowledge of the variations in the lake.
Greetings and good luck!
I agree with Jorge. Homogeneity is very unlikely to be the case. If one fixed location is what you can afford - then probably choose it randomly or based on your best gudgement, the latter to represent a particular biophysical scenario, and then keep observing the same fixed location over time. This will help you get an idea on pattern of temporal changes (using time series modelling) that may be occurring at that fixed location.
I agree with Jorge and Subhash, it is unlikely that your area would be homogenious. You can start by assessing the ecological foot print of your focus area, and that should be the indicator of various potential impacts on the water quality of your area. Then you would be able to make a judgement as to where to concentrate more and and do as Jorge and Suhash have suggested.
If it is to select only one sampling site in such large lake, I think it may be very difficult to generalize the result for the whole lake. However when ever you select the a site, be sure that it should not be very unique than the rest, it shouldn't be a site where direct anthropogenic activities are undertaking, it shouldn't be at a point of discharge (though the objective matters), and is better to follow a transect approach if representative site is to be identified.
Clearly one site is inadequate to characterize a lake the size of Lake Malawi. Halfman 1993 J. Great Lakes Res. 19(3):512-520 has results and a map ff you are interested in examination of temporal changes. Halfman sampled nearshore and deep sites from north to south.
Please check the following:
Whether the sampling point is near outflow portion of then lake
Whether any drinking water station is there
Whether any polluted drain joins the upstream portion of this lake
Whether this point is in the most polluted portion of the lake
Whether this lake portion is the sensitive area of the lake
Elias, I would suggest that the answer depends on the question you are trying to answer. For example, sampling at the Beauchamp location with the same methods Beauchamp used would allow you to make direct comparisons to determine if changes had occurred over time - regardless of whether or not that location is representative of the entire lake. But if you do not intend to make any comparisons to Beauchamp's study or results, you are free to see if there would be a more "representative" sampling location. Ideally you should sample more than one location to be able to capture the spatial variability in the lake. But if you're limited to one location, you should first sample a number of randomly placed locations and then find which of those is closest to the average of all of them in the measures you're interested in.
It depends upon your purpose of sampling. If you are thinking comparing the result with previous work, better follow the same sampling protocol. But you always could have your rational for using some other sampling strategy. I would suggest you better randomize the sample and see if there is homogeneity in terms of physicochemical parameters you consider for measurement.
If you are thinking of looking at the impact of source pollutants, you better consider these as well before developing your sampling protocol. Sampling at the inlet and outlet and comparison of the parameters you consider in these two points gives you rough idea of pollution.
Hello Elias: This is a difficult problem and you have received a lot of good advice. Richard made perhaps the most import point: what question(s) are you attempting to address, and many others agree that water quality in any lake – let alone one as large as Malawi – is highly unlikely to be homogenous. To continue sampling at one location simply because one has data is not warranted. However, that does not mean there is no value in those data; at the very least you could use your ‘new’ samples collected at different locations (along with new samples collected at the historic station) to see how closely patterns or trends in the new data track (are correlated with?) the old.
The problem now becomes where to sample given the resources you have at hand. One way would be to stratify the entire lake by whatever metrics make sense (e.g. proximity to lake inputs, depth, presence of particular habitats, distance along north/south/east/west axes) and then randomly select sites within strata. You could allocate the number based on the expected variability with a type, meaning that each type would be sampled at a different intensity. I expect that this approach will generate a very large number of sites but there are ways to bring the number down to manageable size. You could, for example, use the stratified random scheme but don’t expect to visit all of the sites right away. Step one is to decide how much you can afford ($) and then allocate your resources to a single (or however many you can afford) sample within each type. This approach can get a bit complicated but the point is to account for the characteristic variability of each type.
We undertook a methodological approach to identifying the most appropriate sampling locations considering the circulation attributes of a lake (granted for a smaller lake than in your case). The paper outlining the process can be found here:
https://www.researchgate.net/publication/223043367_Use_of_numerical_models_for_determination_of_best_sampling_locations_for_monitoring_of_large_lakes
Article Use of numerical models for determination of best sampling l...
Hi Elias, if you're interested to check horizontal homogeneities in Lake Malawi in 2002-2012, we can provide maps of about 10 water quality parameters from a free satellite Earth observation database funded by the European Space Agency. I'm attaching a 9-year average of CHL-a as an example. Please be aware that we have rather high confidence in relative gradients, but recommend calibration with in situ measurements to get correct absolute values.
Please check www.diversity2.info and write to [email protected] for access to the full database on our FTP.
By the way, such data is available for about 350 lakes worldwide:
http://goo.gl/9qES0n
Elias - a quick search that includes Hecky and Lake Malawi will provide a number of links to papers regarding a series studies undertaken on Lake Malawi. There is also a summary paper by Bootsma and Jorgensen (contact: [email protected]). I have yet to make it Lake Malawi but it is my understanding the lake has permanaent thermal stratification which is another complicating aspect of the lake.
- Badges
- Science method
- Similar topics
- Biological Science
- Ecology