09/16/05 The Gradient.doc 1/8 Jim Stiles The Univ. of Kansas Dept. of EECS The Gradient Consider the topography of the Earth’s surface. We use contours of constant elevation—called topographic contours—to express on maps (a 2-dimensional graphic) the third dimension of elevation (i.e., surface height). We can infer from these maps the slope of the Earth’s surface, as topographic contours lie closer together where the surface is very steep. Moreover, we can likewise infer the direction of these slopes—a hillside might slope toward the south, or a cliff might drop-off toward the East. From: erg.usgs.gov/isb/pubs/booklets/symbols/reading.html See, this indicates the location of a steep and scary Cliff!09/16/05 The Gradient.doc 2/8 Jim Stiles The Univ. of Kansas Dept. of EECS Thus, the slope of the Earth’s surface has both a magnitude (e.g., flat or steep) and a direction (e.g. toward the north). In other words, the slope of the Earth’s surface is a vector quantity! Thus, the surface slope at every point across some section of the Earth (e.g., Douglas County, Colorado, or North America) must be described by a vector field! A: Yes, there is a very easy way, called the gradient. Say the topography of some small section of the Earth’s surface can be described as a scalar function (),hxy, where h represents the height (elevation) of the Earth at some point denoted by coordinates x and y. E.G.: Q: Sure, but there isn’t any way to calculate this vector field is there? -4-2024x-4-2024y-0.500.51height-4-2024x-0.500.51height-4 -2 0 2 4x-4-2024y09/16/05 The Gradient.doc 3/8 Jim Stiles The Univ. of Kansas Dept. of EECS Now say we take the gradient of scalar field h(x,y). We denote this operation as: ()hr∇ The result of taking the gradient of a scalar field is a vector field, i.e.: ()()hr r∇= A For our example here, taking the gradient of surface elevation h(x,y) results in the following vector field: Q: So just what is this resulting vector field, and how does it relate to scalar field ()hr ?? -4 -2 0 2 4x-4-2024y????09/16/05 The Gradient.doc 4/8 Jim Stiles The Univ. of Kansas Dept. of EECS To see how this vector field relates to the surface height h(x,y), let’s place the vector field on top of the topographic plot: A: That’s right! The gradient of a scalar field provides a vector field that states how the scalar value is changing throughout space—a change that has both a magnitude and direction. -4 -2 0 2 4x-4-2024yQ: It appears that the vector field indicates the slope of the surface topology—both its magnitude and direction!09/16/05 The Gradient.doc 5/8 Jim Stiles The Univ. of Kansas Dept. of EECS It is a bit more “natural” and instructive for our example to examine the opposite of the gradient of h(x,y) (i.e., () ()rhr=−∇A ). In other words, to plot the vectors such that they are pointing in the “downhill” direction. Note these important facts: * The vectors point in the direction of maximum change (i.e., they point straight down the mountain!). * The vectors always point orthogonal to the topographic contours (i.e., the contours of equal surface height). -4 -2 0 2 4x-4-2024y09/16/05 The Gradient.doc 6/8 Jim Stiles The Univ. of Kansas Dept. of EECS Now, it is important to understand that the scalar fields we will consider will not typically describe the height or altitude of anything! Thus, the slope provided by the gradient is more mathematically “abstract”, in the same way we speak about the slope (i.e., derivative) of some curve. For example, consider the relative humidity across the country—a scalar function of position.09/16/05 The Gradient.doc 7/8 Jim Stiles The Univ. of Kansas Dept. of EECS If we travel in some directions, we will find that the humidity quickly changes. But if we travel in other directions, the humidity will change not at all. Q: Say we are located at some point (e.g., Lawrence, KS; Albuquerque, N M; or Ann Arbor, MI), how can we determine the direction where we will experience the greatest change in humidity ?? Also, how can we determine what that change will be ?? A: The answer to both questions is to take the gradient of the scalar field that represents humidity! If ()rg is the scalar field that represents the humidity across the country, then we can form a vector field ()rH by taking the gradient of ()rg: ()()rrg=∇H This vector field indicates the direction of greatest humidity change (i.e., the direction where the derivative is the largest), as well as the magnitude of that change, at every point in the country!09/16/05 The Gradient.doc 8/8 Jim Stiles The Univ. of Kansas Dept. of EECS This is likewise true for any scalar field. The gradient of a scalar field produces a vector field indicating the direction of greatest change (i.e., largest derivative) as well as the magnitude of that change, at every point in space. ()
View Full Document