Geographic information system Study Guide

📖 Core Concepts Geographic Information System (GIS) – a computer system that captures, stores, manages, analyzes, edits, visualizes, and outputs spatial (location) data. Spatial data – describes where a feature is (coordinates, geometry) and its shape (point, line, polygon). Attribute data – non‑spatial characteristics attached to each feature (e.g., population, land‑use type). Map layer – a collection of features that share the same geometry type and attribute schema. Spatial reference – the coordinate system (lat/long, projected) and datum (e.g., NAD‑83, WGS‑84) that give each location a real‑world meaning. Discrete vs. Continuous phenomena – Discrete (houses, roads) are individual objects; continuous (rainfall, temperature) are values that exist everywhere across a surface. --- 📌 Must Remember GIS core functions: store, manage, analyze, edit, output, visualize. Data structures: Raster = grid of cells; Vector = points/lines/polygons. Primary data capture → field measurement, GNSS (GPS), mobile devices, remote‑sensing platforms (satellite, UAV). Secondary data capture → digitizing, scanning, raster‑to‑vector conversion. Coordinate systems: Geographic (lat/long) vs. Projected (e.g., UTM). Datum transforms (Helmert, simple translation) are required when mixing datums. Quality terms: Positional accuracy (how close a coordinate is to true location) and attribute accuracy (correctness of non‑spatial info). Topological relationships: adjacency, containment, proximity – essential for network and overlay analysis. Tobler’s First Law: “Near things are more similar than distant things.” Common GIS software: Esri ArcGIS (commercial), QGIS, GRASS (open‑source). --- 🔄 Key Processes Data Acquisition → Loading Choose primary (GPS, UAV, lidar) or secondary (digitize, scan). Import into a spatial database or file (shapefile, GeoPackage). Data Editing & Topology Check Snap vertices, close gaps, enforce topological rules (no dangles, proper polygon closure). Coordinate Transformation Identify source datum → apply Helmert transformation or translation → re‑project to target CRS. Spatial ETL (Extract‑Transform‑Load) Extract raw layers → clean/standardize attributes → load into unified GIS project. Overlay Analysis (Vector) Union → combine all features & attributes. Intersect → keep only overlapping areas, retain both attribute sets. Symmetric difference → keep non‑overlapping portions of both layers. Raster Overlay (Map Algebra) Apply cell‑by‑cell functions (e.g., weighted sum) to combine rasters into a suitability index. Proximity & Buffer Creation Generate buffer polygons at a given distance; use for impact zones, service areas. Interpolation (Geostatistics) Choose method (IDW, kriging, spline) → create continuous surface → validate with cross‑validation. Geocoding / Reverse Geocoding Geocode: address → coordinate using road‑centerline address ranges. Reverse geocode: coordinate → nearest address number via interpolation on that segment. MCDA (Multi‑Criteria Decision Analysis) Standardize each criterion → assign weights → aggregate (e.g., weighted overlay) → rank alternatives. --- 🔍 Key Comparisons Raster vs. Vector Raster: grid cells, best for continuous surfaces, storage‑intensive for high resolution. Vector: points/lines/polygons, ideal for discrete features, supports topology. Primary vs. Secondary Data Capture Primary: direct measurement (GPS, UAV); higher positional accuracy, higher cost. Secondary: digitizing, scanning existing maps; cheaper, may inherit original errors. Union vs. Intersect Overlay Union: retains all geometry & attributes from both inputs. Intersect: keeps only the overlapping geometry, merging attributes from both. IDW vs. Kriging (Interpolation) IDW: deterministic, distance‑weighted average; simple, no error model. Kriging: stochastic, incorporates spatial autocorrelation; provides prediction variance. Geocoding vs. Reverse Geocoding Geocoding: address → XY coordinate. Reverse: XY coordinate → nearest address (or address range). --- ⚠️ Common Misunderstandings “GIS = map” – GIS is more than mapping; it includes data management, analysis, and modeling. “Higher resolution always better” – Very fine rasters can inflate storage and processing time without improving analysis if the phenomenon’s scale is coarse. “All coordinate systems are interchangeable without loss” – Datum transformations can introduce positional error; always check datum compatibility. “Buffers are always circles” – In projected (planar) CRS buffers are circles; in geographic (lat/long) they become ellipses unless re‑projected. “Interpolation creates real data” – Interpolation estimates values; it inherits the uncertainty of the source points. --- 🧠 Mental Models / Intuition Layer Cake – Think of each GIS layer as a transparent sheet; overlay is stacking them and looking at where colors (attributes) mix. Rubber Sheet Distortion – Projecting the globe onto a flat map stretches some areas; the farther from the projection’s “center”, the more distortion → choose projection that minimizes distortion for your area of interest. Network as Graph – Edges = roads/pipes, nodes = intersections/junctions; classic graph algorithms (shortest path, flow) apply directly. Spatial Autocorrelation Gradient – Visualize a smooth hill: points close together have similar elevations (high autocorrelation). The steeper the hill, the lower the autocorrelation—mirrors Tobler’s First Law. --- 🚩 Exceptions & Edge Cases Datum transformation with only a translation – Works for small regional datasets where ellipsoid differences are negligible; otherwise use a full 7‑parameter Helmert. Absolute vs. Relative accuracy – High relative accuracy (consistent internal geometry) may be sufficient for network analysis even if absolute accuracy is low. Cost distance on non‑continuous surfaces – If the cost surface contains “no‑data” cells, paths cannot cross them; either fill or re‑classify those cells. Geocoding with non‑linear address ranges – Some streets have odd/even numbering or skip numbers; simple proportional interpolation may misplace addresses. --- 📍 When to Use Which | Situation | Recommended Tool / Method | |-----------|---------------------------| | Discrete features with topology needs | Vector data + topological editing (e.g., ArcGIS Geodatabase) | | Continuous surface analysis (e.g., temperature) | Raster DEM/imagery + raster algebra | | Finding service area around a point | Buffer (vector) or Cost‑distance surface (raster) | | Estimating values at unsampled locations | Choose IDW for quick, deterministic; Kriging when spatial autocorrelation is known and error quantification matters | | Integrating multiple criteria for site selection | Weighted overlay (raster) or MCDA with standardized scores | | Large‑scale web map for public consumption | Web Map Service (OGC) + Leaflet/OpenLayers front‑end | | High‑precision field survey | GNSS (RTK) → direct GIS import; ensure datum matches project CRS | | Network routing (roads, utilities) | Geometric network (vector) + network analysis tools | | Temporal trend analysis (e.g., drought over years) | Time‑enabled raster stack → animate or calculate change detection | --- 👀 Patterns to Recognize “Buffer → Clip → Intersect” pattern when you need the portion of a feature that lies within a fixed distance of another feature. “Slope + Aspect → Hillshade → Viewshed” sequence for terrain‑related visibility studies. “High‑density points → Thiessen polygons → Voronoi map” for assigning each location to its nearest point source. “Raster → Reclass → Weighted Sum” pattern for creating a suitability index (common in MCDA). “Large positional error + high‑resolution raster → error propagation” – expect mismatches in overlay results. --- 🗂️ Exam Traps Confusing Union vs. Symmetric Difference – Union keeps all areas; symmetric difference excludes the overlapping area. Assuming all buffers are circular – In a geographic CRS they become distorted; always project to a suitable planar CRS first. Choosing IDW for highly variable data – IDW can create “bullseye” artifacts; kriging or spline may be more appropriate. Ignoring datum when merging layers – Overlays of layers in different datums produce shifted features – a classic “mis‑aligned map” distractor. Treating raster cell size as “resolution” of the phenomenon – Cell size reflects data storage, not necessarily the true variability of the underlying process. Misreading “relative accuracy” as “accurate to ground” – Relative accuracy only guarantees internal consistency, not true‑world positioning. ---