GEOG 400, Advanced GIS, Fall 2020; Week 12 Lecture 1

Transcript

1. GEOG 400: Advanced GIS - Raster Week 12 – Lecture 1 Habitat Modeling & Raster Extraction
2. How we use vector and raster data together to determine what’s good habitat (and kind of good, sort of good, not very good habitat) for different creatures. ...and as part of this, an introduction to extraction analyses. An exercise on modeling habitat suitability for grazing ungulates in southern Utah For this Week… Monday Lecture Wednesday Lecture Lab Using point-based field observations and raster landscape data to measure and map the habitat ranges of monitored coyotes It’s not really possible to only talk about raster data this week. Instead, we’re going to be combining many aspects of rasters and vectors together
3. The Quickest Introduction to Vector Data of All Time Vector Data Types: Point Line (connected points) Polygon (closed loop of points)
4. The Quickest Introduction to Vector Data of All Time Vector Data Types: Point Line (connected points) Polygon (closed loop of points) 1 2 3 4 5 6 Point Number Point Color 1 Red 2 Blue 3 Green 4 Orange 5 Pink 6 Purple Vector Data Attributes:
5. The Quickest Introduction to Vector Data of All Time Vector Data Types: Point Line (connected points) Polygon (closed loop of points) 1 2 3 4 5 6 Point Number Point Color 1 Red 2 Blue 3 Green 4 Orange 5 Pink 6 Purple Vector Data Attributes: Vector Data Formats (that we’ll use): .shp (shapefile) and .csv (text file)
6. A slight detour from GIS data… Rainbow Trout Native Range They’re pretty much everywhere Where do these fish like to live? - In water, obvi - In cold water (50-60 degrees is preferable) - In clear water (relatively little suspended sediment) - In streams with bottoms composed of gravel (16 – 64 mm) - In streams with natural cover (overhanging trees, undercut banks, in-channel wood)
7. Where do these fish like to live? - In water, obvi - In cold water (50-60 degrees is preferable) - In clear water (relatively little suspended sediment) - In streams with bottoms composed of gravel (16 – 64 mm) - In streams with natural cover (overhanging trees, undercut banks, in-channel wood) Where would you want to fish for rainbow trout? Why?
8. Where do these fish like to live? - In water, obvi - In cold water (50-60 degrees is preferable) - In clear water (relatively little suspended sediment) - In streams with bottoms composed of gravel (16 – 64 mm) - In streams with natural cover (overhanging trees, undercut banks, in-channel wood) Where would you want to fish for rainbow trout? Why?
9. Where do these fish like to live? - In water, obvi - In cold water (50-60 degrees is preferable) - In clear water (relatively little suspended sediment) - In streams with bottoms composed of gravel (16 – 64 mm) - In streams with natural cover (overhanging trees, undercut banks, in-channel wood) Where would you want to fish for rainbow trout? Why? Or to put it another way, how much better is the rainbow trout habitat in one of these than another?
10. Where do these fish like to live? - In water, obvi - In cold water (50-60 degrees is preferable) - In clear water (relatively little suspended sediment) - In streams with bottoms composed of gravel (16 – 64 mm) - In streams with natural cover (overhanging trees, undercut banks, in-channel wood) How do we know these things? 1. Catch a whole bunch of fish. 2. Note the sediment size, flow rate, and stream temperature where you caught those fish.
11. And this is where we need to take a brief detour into EXTRACTION. Maybe you’ve (painstakingly) measured depth everywhere you caught a fish. But more likely, you’ve got a map of depth and some GPS points where you caught those fish. You want to extract the values of your bathymetry raster at the fish capture points Let’s go over three very useful extraction methods to get underlying raster data at points and polygons.
12. Extraction Tool 1: Extract Values to Points. I’ve sampled the instances of ponderosa pine trees across a landscape; what’s the elevation of each tree?
13. Extraction Tool 1: Extract Values to Points. I’ve sampled the instances of ponderosa pine trees across a landscape; what’s the elevation of each tree?
14. Extraction Tool 1: Extract Values to Points. I’ve sampled the instances of ponderosa pine trees across a landscape; what’s the elevation of each tree?
15. Extraction Tool 2: Zonal Statistics. I’ve sampled sagebrush patches across the landscape; what’s the average elevation in each of these patches?
16. Extraction Tool 2: Zonal Statistics. I’ve sampled sagebrush patches across the landscape; what’s the average elevation in each of these patches?
17. Extraction Tool 2: Zonal Statistics. I’ve sampled sagebrush patches across the landscape; what’s the average elevation in each of these patches?
18. Extraction Tool 2: Zonal Statistics. I’ve sampled sagebrush patches across the landscape; what’s the average elevation in each of these patches?
19. Extraction Tool 3: Extract by Mask; I have a huge raster but I want to work with something smaller or calculate statistics on a smaller area
20. Extraction Tool 3: Extract by Mask; I have a huge raster but I want to work with something smaller or calculate statistics on a smaller area
21. Extraction Tool 3: Extract by Mask; I have a huge raster but I want to work with something smaller or calculate statistics on a smaller area
22. Where do these fish like to live? - In water, obvi - In cold water (50-60 degrees is preferable) - In clear water (relatively little suspended sediment) - In streams with bottoms composed of gravel (16 – 64 mm) - In streams with natural cover (overhanging trees, undercut banks, in-channel wood) How do we know these things? 1. Catch a whole bunch of fish. 2. Note the sediment size, flow rate, and stream temperature where you caught those fish.
23. 1. Catch a whole bunch of fish. 2. Note the sediment size, flow rate, and stream temperature where you caught those fish. 3. Plot these up as histograms; these are called habitat utilization curves (HUCs); not just for fish
24. But if all you have is bad habitat, you’ll still find fish there – that doesn’t mean they like it. Divide the number of fish in a particular habitat type by the area of that habitat type to get a normalized preference curve. This is also called a habitat suitability index.
25. You’ve done all the work and come up with some habitat suitability curves. Now what? Habitat Suitability Curves
26. You’ve done all the work and come up with some habitat suitability curves. Now what? Habitat Suitability Curves Environmental Variable Rasters
27. You’ve done all the work and come up with some habitat suitability curves. Now what? You’ll get three maps of habitat suitability for each of these inputs. Average them all together… Habitat Suitability Curves Environmental Variable Rasters Habitat Suitability Raster
28. Habitat Suitability Indices are the basis for the most commonly used habitat model out there: USGS Physical Habitat Simulation (PHABSIM) Model Initially developed by USGS (at that point Fish and Wildlife) scientists in Fort Collins in 1978
29. Habitat Suitability Indices are the basis for the most commonly used habitat model out there: USGS Physical Habitat Simulation (PHABSIM) Model Initially developed by USGS (at that point Fish and Wildlife) scientists in Fort Collins in 1978
30. This sort of spatial habitat modeling isn’t without problems, though.
31. This sort of spatial habitat modeling isn’t without problems, though.
32. But even more importantly, do we believe these things? All beavers love this No beaver has ever lived here. Stream Slope Flat Steep DEFINITELY GOOD PROBABLY GOOD PROBABLY BAD DEFINITELY BAD Nature is messy! Remember fuzzy logic from earlier in the semester? CRISP BOUNDARIES FUZZY BOUNDARIES
33. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong
34. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong …but some aren’t so clear.
35. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong …but some aren’t so clear.
36. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong …but some aren’t so clear.
37. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong …but some aren’t so clear.
38. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong …but some aren’t so clear.
39. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs In the class days of the week, some items clearly belong …but some aren’t so clear.
40. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs From the descriptive inputs, we come up with a rule set for determining a crisp (numerical) output
41. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs From the descriptive inputs, we come up with a rule set for determining a crisp (numerical) output
42. FUZZY LOGIC IN 3 MINUTES… A set of rules that produces numerical outputs from descriptive inputs From the descriptive inputs, we come up with a rule set for determining a crisp (numerical) output There’s no reason we can’t do this for modeling habitat suitability, either. If Depth is… And velocity is… Then habitat suitability is… Shallow Slow Nonexistent Shallow Moderate Low Shallow Fast High Moderate Slow Medium Moderate Moderate Medium Moderate Fast High Deep Slow Nonexistent Deep Moderate Poor Deep Fast Medium
43. One example of a fuzzy habitat model… (and yes, there’s an English version!) Defuzzification translates descriptive outputs (good, poor) to crisp outputs (a habitat suitability index of 0.39)
44. This is a class on raster GIS! Why in the world are we covering habitat modeling? Reason 1: relating the occurrence of living things with their habitat almost always involves raster extraction Landcover (LANDFIRE) and climate (PRISM) data are great examples of environmental variables that we’d like to correlate with species occurrence.
45. This is a class on raster GIS! Why in the world are we covering habitat modeling? Reason 2: the outputs of habitat models are almost always rasters Madani et al., 2016 Sarkar et al., 2017 Yost, 2014
46. This is a class on raster GIS! Why in the world are we covering habitat modeling? Reason 3: habitat modeling is one of the most common applied uses of raster data
47. What we’ll do next lecture… What we’ll do in lab… Predicting where cows will graze [image from Macfarlane and others, 2012) Quantifying coyote habitat and range (image from Stout, 2016)
48. For Lab this Week… Read pages 547-561 on core area mapping (we’ll use this for coyote habitat modeling)