In my field data the main source of uncertainty is spatial, I have camera trap and village points and a shape file of rivers. In case of the points, the location where they’re located in qgis may not be the most accurate, so estimations of the the distances between them and the rivers have uncertanity.
Water data, specifically groundwater data, can use satellite imagery but more often than not has to ground-truth because of high levels of uncertainty. Satellite imagery can help detect ground-water potential areas but it’s different to determine groundwater bc of complex and varying geological features, fracture systems, and flow rates. All three uncertainties exist: thematic (boundaries of groundwater), spatial (exact location), and temporal (groundwater flows). This is inevitable due to the nature of the water. Generalizations have to made, mostly because the nature of groundwater locating via satellite is to determine potential and not exact accuracy and then ground-truthing has to corroborate the potential of groundwater levels, quality, and location. Making maps become a complex task with groundwater because of the need for both theoretical, field and experimental data to extrapolate and make a lower resolution map that works. More advanced technology is underway to detect groundwater levels, quality, flow, and location but the accuracy of groundwater detection and montiroing hasn’t been research a lot, there needs to be more research into data accuracy.
Uncertainty: I worked with a forest Reserve in monitoring biodiversity. I find out that satellite image do not take into account farm areas as open areas.
Also recording the location of wildlife might not be accurate as we do not track the animals all day and night. Uncertainty is inevitable, for map making or making management decision the degree of uncertainty should be reported.
This really took me back to my undergraduate project titled “abundance of soil arthropods in the botanical garden Kogi state University Anyigba, Nigeria”
The idea of uncertainty truly exist!.
I personally encountered several uncertainties in project emanating from different human arthropogenic activities.
Just like TEMPORAL UNCERTAINTIES
Bush burning
Large animal passing through my pitfall
traps
Animal grazing through the garden etc.
RESULT OF THOSE UNCERTAINTIES
The project took more time than
required
It cost more than expected
Time wasted etc.
RECOMMENDATION
Different sampling data to take the mean is necessary to reduce uncertainties.
Time of sampling is also considered to avoid something that changes seasonally WEATHER CONDITION
Good standard software for more of digital data.
Thanks
incorporating GIS in a certain project is something exciting and challenging. Yes, there is certainly issues around the way satellite data is collected, processed and made available to the users and it depends on what source of data is being used. Like some of the areas were satellite imagery are not really clear, sometimes they are dependent with geographical locations.
In regards to satellite imagery, there could be temporal uncertainty (i.e. vegetation cover can change over time and weather patterns change frequently). With weather, cloud coverage can obscure the land cover details below (at least in other aerial imagery that I’ve worked with). I would say that these factors are inevitable as they are not easy to control for. You can make some adjustments by looking at satellite imagery changes over time to adjust for changes in vegetation cover or by working with the most up to date imagery of high/usable quality. This would result in some simplification of real world circumstances. I would think that this would result in some uncertainty in your analysis/map-making/recommendations, but by using a combination of sources over a long period of time (bigger data set) you can try to reduce the uncertainty in your data.
There are quite several type of uncertainty that I have come across in my field examples include, issues of weather despite planning ahead and confirming with the weather at times it can be tricky and can result in changes of sampling to later dates. Sometimes decoding the end of a particular habitat type can also cause some challenges as to the demarcation of a clear boundary. With humans as well I have seen that there could be error in the entrance of quantitative data as different research assistants are used in different sampling intervals at times
I think uncertainty is indeed inevitable, mainly due to changing times and rapid changes on the ground itself. Temporal uncertainty coupled with lack of human skills in such a way that field data snapshots may be unclear and not truly representing what`s on the ground. I suggest maybe having several researchers on the ground doing the same work and an assessment be done by an independent evaluator to check on possible limitations.
Cross-verification of data is certainly one way to identify un-certainty (may be not to irradticate it but at least to ‘quantify’ it). However, this must be balanced by the extra cost and time involved. In real world projects, it is often about identifying what an “acceptable” level of uncertainy is and working with that. Trying to “iradicate uncertainty” can dramatically increase data collection/processing costs in a project and make it unviable, therefore knowing what is acceptable for sucessful project implimentation is a good pragmatic approach to take.
Because I work in a remote area, there is a lot of error associated with any gps coordinates we take. So much of the spatial data I collect is a single point with high accuracy and then measuring the distance in meters and the heading of the animals movements to try and limit error.
I am a beginner but I will try answer.
Relating to my project (it is based in a rural area in South Africa), when I want to obtain spatial data from satellite imagery I am unable to do so due to the lack of information available. I am unable to zoom into some parts of my study site (i.e. open street view even on google earth pro).
I hope I did not go off topic, I am still learning the terminology. 
Uncertainty in my GIS work can result from getting accurate data to represent vegetation cover. Unless there are continuous data for vegetation change
Most of the time, the uncertainty that exists with the data I collect is related to weather conditions. For example, if I am conducting field sampling, the weather plays a large role in when it can be done. While I would like to conduct it on a set schedule, it must be varied some times. Also, as some have mentioned earlier, depending on satellite data also is influenced by the weather. If there is heave cloud cover, then the data is of poor quality or may not be available. All of this effects the spatial resolution of the data set.
I completely agree with you… this is one of my major challenge because it is very difficult or almost impossible going back to the same exact location you took your previous samples from.
Kroem, clouds can be corrected using some atmospheric correction methods and possible increasing your sample size and sample duration. Most importantly you must ensure the weather is monitored and data is collected on a non-cloudy day.
Personally my uncertainties can be both temporal and spatial. Temporal in the sense that samples may not be collected on the same day monthly or annually due to some environmental conditions like Rivers state where it rains every day during the rainy season.
Secondly, working with satellite data is one of the most difficult area because of precision like everyone has mentioned; using the right satellite for the environment, issues with mixed pixel especially when working on a lake and heavy cloud interference. For the cloud interference, I handled it using different atmospheric correction methods.
But i will like a better way to go about because i am very interested in satellite remote sensing of the aquatic environment.
Mapping in rural areas is often a problem and it may require trying different mapping sources to identify the one most relevant to your area. There shouldnt be a problem of “Zooming in” on any area, but the issue is the level of detail and resolution of imagery you get when you do “zoom in”.
If freely available information is not available of required quality, it may be that “paid for” commercial data is available that gives you the level of detail or resolution you need.
If no data is available and is essential for your project, then its a case of commisioning “primary data capture” (eg. on site survey, use of Drones to capture site specific imagery, aerial imagery from commisioned aeroplane flight, or even tasking a satellite to capture imagery of a given place at a given time - expensive but possible if its an essential requirement of a GIS project!).
1 Like
Uncertainty can come from:
- Human Errors
- Environmental Characteristics
- Instrument Errors
Thank you so much for your comment 
I agree that uncertainty in temporal information would affect the judgment on the events, which would lead to inaccurate or wrong findings if we used an event with uncertain time as an evidence or other analysis resources. It is also said that we have temporal information from GPS logs and transaction data. On the other hand, spatial information systems are designed to manage large amounts of information about geographical objects representing real-world phenomena. Every set of spatial data based on measured data is susceptible to uncertainty, which is expressed in many forms. Should the uncertainty be ignored, results of analyses will remain logical, but the research will yield inaccurate or even misleading results. Confidence in an information system that ignores data uncertainty is, therefore, undermined. Hence, the widespread awareness of data uncertainty will be of vital importance to all spatial information system users