I will say Vector data is what we mostly collect in the field. During our recce transect surveys, and wildlife camera trap. We collect for example points of animal dug, footprints, calls and even direct sightings etc. hunters camps too.
When we observe say the dung of a Blue duiker, a range of information is collected including the canopy cover and leaf cover.
However at the start of a survey day and at any survey point, we do record some covariates like weather condition. Am thinking raster data. Am I correct?
In my case, I need both raster and vectors. My urbanisation cathegories, green space/road/builiding/other concrete are vector polygons. My sampling points are vector points. I also need raster for climate data which are continuous. But to make the polygons I would need an Open map layer what is a thematic raster.
In many environmental GIS applications, there is often ambiquity between how different datasets should be handled. Vegetation classes/regions are often represented as Vector data defining discrete polygons, but reality is often different with their being a ‘transistion zone’ between different habitat/vegetation types rather than a definite “hard line” between the the two classes. Another example is topographic infoirmation. It may be repsented from survey data as height “points” but these are only sample locations from a continously variable “surface” which can then be interpolated from the “points” to generate a continuous surface represented as a “raster” which can then be converted to contours which are held as a “Vector” - three different formats to represent a single physical characteristic! - yes its complicated
From a technical data structure, Raster data is a grid format, comprised of “pixels” (a row of pixels known as a 'raster) so an aerial photograph, scanned map, surface height model etc are stored as a Raster.
A Vector is a series of “verticies” (coordinate points) which make up a line (or vector). If the first and last vertices are the same, then the vector will enclose an area and is known as a “polygon”. So a vector can represent a Road, river, survey transect , or as a polygon - land parcel, animal territories, lakes etc.
There can be a lot of overlap in the way the data structures are used so a river can be represented as a vector line (say, the centre line of the river), but it can also be represented as a polygon (with a line running down the left bank, across the river, then up the right bank to form a long thin polygon, or as a raster with pixels coloured “blue” repsenting the watercourse.
For my data on nest box success, I think that I would use vector line data to represent data at each nest box site in the field and these would be points instead of lines or polygons. I could also use the raster data model to model the different habitats at each nest site and I think that this would be a thematic raster model instead of a continuous raster model.
It is a little ambiguous because there is so much data at each nest site to consider so I could see it being a difficult to choose between points or lines to connect the data of all the nest sites. However, since the nest box sites are separate from one another, I think points would be the best choice. There is less ambiguity when considering the habitat type because that is a definite theme instead of a continuation.
I guess my questions would be how to implement different data collected at each site into the representation such as species, clutch size, nest box material chosen, temperature etc. Would this have to be represented in separate maps for each type of data to focus on the specific data?
I hope to one day be collecting rare plant species data, which would include points/locations (vector) of identified rare species or polygons of the range/small area in which that species is found. Other data would include environmental (raster) data, including: year of the most recent fire, rainfall, slope, aspect, etc. I could see selecting which data model to use could get tricky. For example, when it comes to my data example, you could either make polygons of the years of fire (vector) or make it a raster with pixels identifying the year of the most recent fire.
A question I have is: if I’m surveying vegetation cover in selected quadrats, how would I turn that into a continuous vegetation cover raster over the entire study area?
For my master’s project, I collected vector data (point locations of bat sightings). But then I used a variety of different rasters (environmental data that were freely available) to input everything into a species distribution model. This included mean annual temperature, distance to the nearest river, distance to the nearest forest edge, slope, elevation, vegetation type, predator control, etc. Some rasters were continuous, such as distance to the nearest river/forest, slope, and elevation. Then, some were thematic (categorical) like vegetation type which was sorted into 10 different categories, and predator control which was split into 1- yes or 0- no.
Although I ended up using a raster form of this data in the end, often times this data started out as vector data. For example, distance to the nearest river started out as a vector (lines) of all the rivers in the area. Predator control is an example of data that I could easily represent with vector or raster files- they are polygons of an aerial 1080 drop that I had to convert to a raster of 1’s and 0’s for my manipulation, but I think either way is valid.
I may use both vector and Raster spatial data modeling to represent my field data. Locations of ( our beneficiaries’ homes, beehives & grass plots) would be represented by vector by points and for size of grass plots vector by polygons would be suitable model.
for elevations of beehives; locations and distances of the beehives to boundaries of the wolf habitat in the area, the appropriate model would be Raster & the layers be continuous.
Question- For the field survey data below I am not sure/confident if it would be represented by " vector by lines" or " Raster by continuous" data modeling. I would appreciate for the answer with justification. Thank you.
The field survey data is " the distance in meters beehive is located from the nearest boundary of the wolf habitat area" ,
I am recording camera trap locations which would be vector / points and also polygons for illegal land clearings. I think maybe i could use rasters to class land by land uses (thematic) and this could also be time-based (changing over time as more forest is sadlly cleared). So many uses!!
HI Cee, if you are trying to represent sampling “points” then the format you will need is as a Vector point data rather than linear data (ie. you have a single point represented by a single co-ordinate).
If you wanted to represent a survey transect or route you took to take the samples, then this would use a “linear” vector format (a line represented by a number of ‘vertices’ which have co-ordinates for each segment of your route going from point A to Point B to Point C etc).
so then the next question - do you use vector / pologon points to define the information in the rasters… or rather little tiles that form the raster dataset? (not sure if i’m using the correct grammar there!!
For mine, squirrel sightings are vector points and habitat would be raster thematic as it is broadly blocked into habitat types. I am a little bit confused however on why an aerial photo would come up continuous raster - can anyone shed some light? Thanks
Im currently collecting data on different habitat types and areas of degradation on privately owned farms and wildlife ranches. I would use a thematic raster model to map the different habitat types and a polygon vector model for mapping areas of degradation
I’m afraid I don’t have any data of my own yet as I’m totally new to this line of work. I will see what data I can get hold of.
I’m looking at going into a marine biology field in particular so if anyone can suggest a good place for me to get any relevant data I’d really appreciate it. Thank you
I do nesting turtle survey so vector data collection can works for the gps point were the turtle nest are. we can draw up the hot spots at the beach represented by polygons. and in the case of erosion I guess can be represented by raster.
In my project, I collect vector data (point locations of count station) and the perpendicular distance of me to my study subject for getting the detection probability. Then, I will also try to use a variety of raster data (vegetation type, density and cover) into density surface modelling where some of them are thematic raster (vegetation type) and continuous raster (density and covers). I guess I will end up using continuous raster data (detection probability of birds in accounts of the raster data of my environmental variable) in the end. It is still a little but confusing to me, just trying to make sense of my data and different data type.
