Uncertainty in my work might be due to GPS system and it’s units used, camera trap data will also introduce uncertainty as well as direct observation transects. One way to save this might be taking a lot of data and have a solid or strong methodic way to take that data, also come back several times as long as it’s possible.
I don’t know yet what will come out of my study and what I’ll face when I’m on field. But I hope I can save errors and uncertainties and have a good control of my variables and methods.
Uncertainty in the field when tracking Texas horned lizards comes in many forms and we do our best to minimize them. One such spatial uncertainty arises from issues with GPS equipment synchronizing with satellites. The more satellites the GPS units are able to use, the more precise the location recorded will be. Sometimes there just aren’t enough satellites, causing data points to potentially be off by a few feet. Using survey grade equipment, or only recording locations on days when there are enough satellites synchronized would definitely decrease the amount of uncertainty.
Another uncertainty was temporal. We started with an extremely small population of translocated and resident lizards in the study area, so the data could be skewed until more individuals were found or added by translocation. Because of this, early data would not be included in the study until a large enough population could be tracked to account for any sampling bias.
1 Like
Uncertainty i will face in my field work are ocean current,rainy,turbidity and cloudy in coral reef assesment(which are both spatial and temporal)
Part of the uncertainty in satellites can come from a lack of precision, they have clouds which make measurement difficult.
Best regards,
PHON Kroem
2 Likes
I feel the same as I am also quite new (or revising) and also do not have any current projects! Thank you for sharing!
Hi everyone! I must admit that I am struggling to think of examples as I do not have any current projects and I am not currently working but hoping to be able to use GIS in future work. I guess I can imagine that if I was managing a site there would be spatial uncertainty due to vegetation or habitat boundaries not being quite clear cut in terms of the species present in them. There would also be temporal uncertainty due to weather and condition changes. If I was focussing on vegetation cover I guess I would try and do a survey in each season to determine the habitat types to reduce uncertainty. This is all depends on time constraints though!
I think that it is a bit hard to map nature generally as it varies so much and does not contain straight lines! But we still need to represent it to the decision makers so I guess we can do our best to do this but it will never be what is truly happening on the ground! Maybe that is something we just need to accept!
2 Likes
The type of uncertainty in my area is Spatial. We create density and heat maps to show animals ranging areas. The biggest challenge to determine the search radius so that the exact home range area can be calculated. This can be partly solved by collecting as many location points as possible so one can have a rough idea on where the animal roams.
The uncertainty in my data will be temporal uncertainty. This is because the data needed for my research are thermoregulation and movement data. Since I will be recording data of more than one lizards in several locations at the same time of day. This causes uncertainty as the weather is inevitably not the same for every minute and generally, the weather could influence the thermoregulation data, movement, and also animal behavior. Ways to reduce the uncertainty are maybe by adding a weather column or description and also which part of the day it is collected. Besides, we can reduce uncertainty by recording data during the active period of animals and also avoid collecting data during unpleasant weather. We can also try recording data in one location for a day rather than several locations in a day. Anyways we cannot avoid the uncertainty fully but reducing it will make our data better.
2 Likes
I am using an environmental dataset that is generated for across the globe, so it is a fairly course-grained dataset so there will be some spatial uncertainty.
I have also had a problem where I have been trying to work out the shortest distance between two populations (marked as geographic points) through the ocean, as I work on a marine species. I tried to use a program in R called Marmap, which has a function to calculate shortest distance through the ocean but it works by using bathymetric data - so you can set an upper limit on the depth, above which is excluded as a possible path - but the species I’m looking is at a very shallow depth/is coastal so many of the locations for which I have values are above sea level because of the uncertainty of the spatial dataset. So I haven’t quite got it to work yet.
When observing elephants in the field we take a distance from us (on the road) to where the elephants are to try and deal with inaccuracy (uncertainty) of our GPS location, with the idea of then adapting our readings to get vegetation type from vegetation maps.
One of the uncertainty in my data is the confusing boundary of the forest vegetation especially for savannah vegetation. I am always not sure where the boundary stop and to draw a clear line has always been a challenge.
In the species survey data I collect there will definitely be uncertainty brought about my own knowledge and ability to correctly identify target species. This can be reduced to a certain degree by researching/training prior to carrying out the survey, but some uncertainty will only be resolved by experience gained over years.
There has also been spatial uncertainty in trying to locate previous survey locations after almost a decade has passed and vegetation and location characteristics have altered. This will be improved going forward as I am now adding GPS coordinates to the survey locations whereas previously there was only descriptive text.
There is also a measure to reduce temporal uncertainty by only conducting the surveys at the certain time of year and when suitable weather conditions are present.
One uncertainty that I’ve noticed is a few of my datasets that I’m working with is a boundary for lake varies a bit. This specific lake doesn’t affect my work too much, but skewed boundaries can alter results.
One major source of uncertainty for the survey work I complete could be the time of day I visit the raven nests. This temporal uncertainty might result me catching a bird perched near nest in the morning or not viewing them because of foraging or other behaviors. We could increase our temporal resolution by taking more surveys throughout the day to get a better picture of how many chicks are still around the nest site, and if the adults are with them or not.
I’m not sure if I completely understand this concept yet, but when I helped with frog surveys last year, a limitation of that data was that surveys were only being conducted twice a year so I think there is likely to be temporal uncertainty with changes in population size that may not be recorded? This was in a tropical environment so seasonal changes are not so uncertain however in a temperate environment I think sampling more regularly is a way to reduce uncertainty as you gain more information about fluctuations and patterns over the different seasons.
My data could contain uncertainty of the spatial type. They could have been obtained better if a GPS was used when collecting the specimens to have more precise information on the locations.
In my project, we always faced spatial uncertainty especially when you are using data from GBIF. The incorrect coordinates can be inevitable when you have millions of data in your database and this will affect on the result. Currently, we are checking the data point one by one and delete the uncertainty points.
The main uncertainty in my geospatial data is in the vegetation cover. Layers are often out of date for years (temporary uncertainty). Another problem with these layers is that many times there are more types of vegetation than we need, so we are forced to simplify them for analysis. I honestly don’t know of a way to reduce this uncertainty … Any ideas?
We have uncertainty in determining if the boundaries of seagrass meadows are changing over time. One reason is the inaccuracy of the GPS readings at +/- 2m (at best), and the other is the natural shift of sand and sediment when the winds change between the dry and rainy seasons.