Geographical Techniques

Overview

Geographic information is information about places, with a location component and an attribute component. Location answers where something is, in coordinates or in relation to other places; attributes answer what it is, such as population, rainfall, soil type, or land use. Data can be spatial, showing the geometry of points, lines, and areas, or non-spatial, showing tables of attributes linked to those geometries. Data quality matters more than data quantity: accuracy, precision, completeness, currency, and consistency decide whether an analysis is trustworthy. A map made from outdated, incomplete, or inaccurately located data will mislead, however sophisticated its presentation. The geographer's first task with any dataset is therefore to ask where it came from, when it was collected, at what scale, and with what known errors, before any analysis begins.

Geographic information and map making

A map is a selective, symbolised representation of the Earth, or part of it, on a flat surface, reduced in scale and generalised for a purpose. Scale is the ratio between distance on the map and distance on the ground, expressed as a representative fraction, a statement, or a graphic bar scale. Large-scale maps show small areas in detail; small-scale maps show large areas with less detail. The choice of scale is a choice about what can be shown: a large-scale map can show individual buildings and field boundaries, while a small-scale map can show national patterns but must omit local detail. Generalisation is the necessary simplification that follows from scale reduction: features are simplified, displaced, exaggerated, or omitted, and a good map generalises deliberately for its purpose, not carelessly.

Projections are methods of transferring the curved Earth onto a flat map, and every projection distorts something, because a sphere cannot be flattened without distortion. Projections preserve different properties: conformal projections preserve shape locally but distort area; equal-area projections preserve area but distort shape; equidistant projections preserve distance from a point or along certain lines but distort other properties. The choice of projection should match the map's purpose: an equal-area projection for comparing the sizes of countries or regions, a conformal projection for navigation where direction and local shape matter, and a compromise projection for general world maps where no single property is paramount. Map symbolisation uses points for discrete locations, lines for routes and boundaries, and areas for regions, with visual variables — size, shape, colour, value, texture, and orientation — to encode quantity and quality. Choropleth maps shade areas by a statistical value, and they must use rates or densities, not raw totals, or large areas will dominate misleadingly; dot maps show distribution by placing one dot per unit quantity; proportional symbol maps scale symbols by value; and isoline maps join points of equal value, such as contours for height or isotherms for temperature. A map without a title, scale, legend, north arrow, and source is incomplete, because the reader cannot judge what it shows or how far to trust it.

GIS and remote sensing

A geographic information system is a computer system for capturing, storing, manipulating, analysing, and displaying spatial data. Its power lies in layers: different datasets — roads, rivers, soils, population, land use — are stored as separate layers in a common coordinate system, so they can be overlaid, queried, and combined. GIS data are stored in two main models. The vector model represents features as points, lines, and polygons with precise coordinates, suited to discrete features such as roads and administrative boundaries. The raster model represents space as a grid of cells, each with a value, suited to continuous surfaces such as elevation, temperature, and satellite imagery. GIS analysis includes buffering, creating zones around features to show proximity; overlay, combining layers to find areas meeting several criteria, such as land suitable for a purpose; network analysis, finding shortest routes and service areas along roads; and spatial querying, selecting features by location or attribute. GIS does not create truth by itself: its output depends on the quality, scale, and compatibility of its input layers, and overlaying layers of different scales or dates can produce spurious precision, where the output looks more accurate than its inputs justify.

Remote sensing is the acquisition of information about the Earth's surface without direct contact, mainly by sensors on satellites and aircraft that record reflected and emitted electromagnetic radiation. Passive sensors record sunlight reflected from the surface and thermal radiation emitted by it; active sensors, such as radar, emit their own energy and record its return, allowing imaging at night and through cloud. Different surfaces reflect radiation differently across wavelengths, producing spectral signatures that allow identification of vegetation, water, soil, and built surfaces. Resolution has several dimensions: spatial resolution, the size of the smallest feature that can be distinguished; spectral resolution, the number and narrowness of wavelength bands; temporal resolution, how often the same area is revisited; and radiometric resolution, the sensitivity to differences in brightness. Image interpretation uses tone, colour, size, shape, texture, pattern, shadow, and association to identify features, and it is increasingly supplemented by digital classification, in which computers assign each pixel to a land-cover class by its spectral signature, supervised by training samples or unsupervised by statistical clustering. Remote sensing is invaluable for mapping large, inaccessible, or rapidly changing areas, for monitoring deforestation, urban growth, floods, and crop condition, but it requires ground truth — field checking — to validate classifications, and it cannot directly measure many social variables, such as income or opinion, that must be inferred indirectly or collected by other means.

GPS

The global positioning system is a satellite-based navigation system that provides location, time, and velocity anywhere with a clear view of the sky. It consists of three segments. The space segment is a constellation of satellites orbiting the Earth, each broadcasting signals that include its position and the time the signal was sent. The control segment is a network of ground stations that monitor the satellites, predict their orbits, and upload corrections. The user segment is the receivers carried by people, vehicles, and instruments, which calculate position by trilateration: measuring the distance to at least four satellites from the travel time of their signals, and intersecting the resulting spheres to find the receiver's position in three dimensions, with the fourth satellite resolving the receiver's clock error. The principle is simple, but accuracy depends on several factors: satellite geometry, expressed as dilution of precision, where widely spread satellites give better positions than clustered ones; atmospheric delay, as signals slow in the ionosphere and troposphere; multipath, where signals reflect off buildings or cliffs before reaching the receiver; and clock errors. Differential GPS improves accuracy by using a fixed base station at a known location to calculate corrections and apply them to nearby receivers, which is essential for surveying and precise mapping. GPS has transformed fieldwork, allowing geographers to record locations, tracks, and sample points accurately and to navigate to remote sites, but it does not replace map reading, field observation, or local knowledge, and it fails where sky view is blocked, such as in deep valleys, dense forest, or urban canyons, and where its signals are deliberately degraded or jammed.

Quantitative techniques

Quantitative techniques give geography the tools to describe patterns, compare distributions, and test relationships. Central tendency summarises a distribution in a single value: the mean, the arithmetic average, sensitive to extreme values; the median, the middle value when data are ordered, resistant to extremes; and the mode, the most frequent value, useful for categorical data. Dispersion describes spread: range, the difference between highest and lowest values, simple but sensitive to extremes; variance and standard deviation, measuring average squared and average distance from the mean, which are the standard measures of spread for quantitative data; and the coefficient of variation, which expresses standard deviation as a percentage of the mean, allowing comparison of variability between datasets with different means. A distribution can also be described by its shape: symmetrical, skewed to one side by extreme values, or bimodal with two peaks. Choosing the right summary matters: a mean income can be misleading where a few very high incomes pull the average up, while the median shows the typical person better.

Correlation measures the strength and direction of a linear relationship between two variables, expressed by the correlation coefficient, which ranges from minus one for a perfect negative relationship, through zero for no linear relationship, to plus one for a perfect positive relationship. A high correlation does not prove causation: two variables may both be caused by a third variable, or the relationship may be coincidental, and a scatter diagram should always be examined before interpreting a coefficient, since a strong non-linear relationship can have a low linear correlation. Regression extends correlation by fitting a line that predicts one variable from another, allowing estimation and forecasting, with residuals, the differences between observed and predicted values, showing where the model fits well and where it fails. Spatial statistics add a geographical dimension: nearest-neighbour analysis tests whether a point pattern is clustered, random, or dispersed by comparing observed distances between points with those expected under randomness; and measures of spatial autocorrelation test whether nearby places tend to have similar values, which is common in geography because climate, soil, and social processes vary smoothly across space. Statistical testing, through significance levels and confidence intervals, guards against mistaking a chance pattern for a real one, but it does not replace substantive interpretation: a statistically significant result can still be geographically trivial, and a non-significant result can still matter if the sample is small.

Morphometric analysis

Morphometric analysis is the quantitative measurement of landforms, above all drainage basins, to describe their form and to infer the processes shaping them. A drainage basin is the area drained by a river and its tributaries, bounded by a watershed. Its basic measures include area, perimeter, length, and relief, the difference between highest and lowest elevation. Stream order, most commonly by the Strahler system, classifies streams by their position in the network: headwater streams with no tributaries are first order; where two first-order streams join, a second-order stream forms; where two second-order streams join, a third-order stream forms, and so on, with a lower-order stream joining a higher-order one not increasing its order. The number of streams of each order typically decreases geometrically as order increases, a relationship known as the law of stream numbers, while average stream length increases with order. Drainage density, total stream length divided by basin area, measures how dissected the landscape is: high drainage density indicates impermeable rock, steep slopes, sparse vegetation, or heavy rainfall, all of which favour surface runoff and channel formation; low drainage density indicates permeable rock, gentle slopes, dense vegetation, or low rainfall, favouring infiltration over runoff.

Other morphometric measures include stream frequency, the number of streams per unit area, which complements drainage density; bifurcation ratio, the ratio of the number of streams of one order to the number of the next higher order, which reflects geological structure and tends to be low in uniform, undisturbed terrain and high where structure controls the network; form factor and circularity ratio, which compare basin shape with a circle or with its length, distinguishing elongated basins, which produce delayed, attenuated flood peaks, from circular basins, which concentrate runoff and produce sharper flood peaks; and relief ratio and ruggedness, which combine relief with basin dimensions to describe steepness and erosion potential. Hypsometric analysis, plotting the proportion of basin area above each elevation, shows the stage of erosion: a convex curve indicates a youthful, little-eroded basin, an S-shaped curve indicates maturity, and a concave curve indicates old age and extensive erosion. Morphometric analysis is powerful because it turns a map into numbers that can be compared between basins, but it must be interpreted with geology, climate, and land use in mind: similar numbers can arise from different combinations of causes, and numbers alone, without field observation, cannot prove a process.

Vector and raster compared

Map projections compared