Interpolation principles lecture (NCSU Geospatial Modeling and Analysis)

1. Caption evidence
   Observation

   Slide text reads “Bivariate interpolation: find surface representation based on values measured at 2D points”.

2. Audio
   Observation

   The lecturer explains using interpolation to convert scattered points, profiles, or isolines to raster and to compute denser or smoother TIN or higher-resolution raster.

Bivariate interpolation

Interpolation that finds a surface representation from values measured at 2D points.

2D point data used to construct a surface or raster representation

1. Caption evidence
   Observation

   Slide text reads “Trivariate: values measured at 3D points to 3D raster”.

2. Audio
   Observation

   The lecturer says interpolation can convert values measured at three-dimensional points to three-dimensional raster.

Trivariate

Interpolation from values measured at 3D points to a 3D raster.

3D point data used to construct a 3D volume model

1. Caption evidence
   Observation

   Slide text reads “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)”.

2. Audio
   Observation

   The lecturer introduces higher-dimension interpolation for points in three dimensions over time.

1. The spoken explanation is cut off before the lecturer completes the quadrivariate example.

Quadrivariate

Interpolation from values measured at 3D points over time to a 4D raster combining 3D space and time.

Spatiotemporal point data with three spatial dimensions plus time

1. Caption evidence
   Observation

   Slide text reads “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)”.

2. Audio
   Observation

   The lecturer says that when values are measured over time in three-dimensional space, quadrivariate interpolation is needed because there are four variables.

Quadrivariate

An interpolation setting with four variables, namely three spatial coordinates plus time.

Values measured at 3D points over time; output described as a 4D raster.

1. Caption evidence
   Observation

   Slide text states “4D raster (3D space + time)”.

2. Audio
   Observation

   The lecturer describes obtaining an actual four-dimensional model by interpolating in 3D space and time.

4D raster

A gridded representation whose dimensions are three spatial dimensions plus time.

Used for the groundwater pollution example over 10 years.

1. Audio
   Observation

   The lecturer says each point has its x, y location.

2. Diagram
   Observation

   Left visualization shows surface points distributed over a horizontal plane.

x, y

Horizontal spatial coordinates of a measurement point.

Spatial location in the groundwater example.

1. Audio
   Observation

   The lecturer says the z depth is exaggerated in the 3D-space view.

2. Diagram
   Observation

   Right visualization is labeled “points in3D space” and shows vertical dispersion of points.

z

Vertical coordinate or depth in the three-dimensional spatial view.

Depth coordinate in the 3D-space visualization.

1. Audio
   Observation

   The lecturer says that in the time-distribution view the z value does not represent the actual three-dimensional location but rather distribution in time.

2. Diagram
   Observation

   Left visualization is labeled “surface points in time” and includes a marked length labeled “250 days”.

1. The lecturer’s wording uses z for the plotted vertical axis in the time view; whether this is a formal coordinate symbol or a descriptive reference to the display axis is not separately defined on screen.

z

In the left “surface points in time” visualization, the plotted vertical extent represents time rather than physical depth.

Time-distribution visualization of well data.

1. Diagram
   Observation

   A marked length in the left visualization is labeled “250 days”.

2. Audio
   Observation

   The lecturer says, “for example this length here is 250 days,” then says another point would be around 500 days.

1. The exact numeric label for the second point is spoken rather than clearly shown as a separate printed value.

250 days

Example time length used to illustrate how vertical extent encodes elapsed time in the surface-points-in-time view.

Temporal scale in the groundwater well example.

1. Formula
   Observation

   Bivariate formulation slide states “Given m-points ($x_i$, $y_i$, $z_i$), $i=1$,m”.

2. Audio
   Observation

   The lecturer says, “So we have m points given by their coordinates.”

m

Number of given discrete points.

Index range $i=1$,m in the bivariate interpolation formulation.

1. Formula
   Observation

   Slide shows “($x_i$, $y_i$, $z_i$), $i=1$,m”.

2. Audio
   Observation

   The lecturer says they will look at the bivariate formulation and have m points x, y, z.

($x_i$, $y_i$, $z_i$)

Coordinates of the i-th given discrete point, with $x_i$ and $y_i$ as independent spatial coordinates and $z_i$ as the observed value.

$i=1$,m in the bivariate formulation.

1. Formula
   Observation

   Slide states “find such $F(x,y)$ that for each $i=1$,m”.

2. Audio
   Observation

   The lecturer says they are trying to find a function that passes through these points.

$F(x,y)$

The interpolation or approximation function to be found from the discrete data.

Function of two independent variables x and y in the bivariate formulation.

1. Audio
   Observation

   The lecturer states that the lecture discusses spatial interpolation and approximation.

2. Caption evidence
   Observation

   Title slide reads “Spatial interpolation and approximation section I” and “Geospatial Analysis and Modeling: Lecture notes Helena Mitasova, NCSU MEAS”.

3. Caption evidence
   Observation

   Outline slide lists “spatial interpolation: definition, applications”, “methods and their properties”, “optimizing interpolation using splines”, “evaluating interpolation accuracy”, and “trivariate interpolation of volumes and topo-climatology”.

This opening segment defines the lecture topic as spatial interpolation and approximation and previews the structure: first the definition and applications of spatial interpolation, then methods and their properties including Voronoi/TIN, IDW, geostatistics, radial basis functions and splines, followed by optimization, accuracy evaluation, and trivariate interpolation of volumes and topo-climatology.

1. Applies to the introductory overview of the lecture rather than to a worked mathematical derivation.

1. Caption evidence
   Observation

   Slide text reads “Bivariate interpolation: find surface representation based on values measured at 2D points”.

2. Audio
   Observation

   The lecturer says interpolation is used to convert scattered points, profiles, or isolines to raster, and to compute denser or smoother TIN or resample continuous-data raster such as elevation to higher resolution.

Bivariate interpolation is presented as the construction of a surface representation from values measured at 2D points. The video gives three application patterns: converting scattered point sets, profiles, or isolines into raster form; computing a denser or smoother TIN; and resampling a raster representing continuous data such as elevation to higher resolution.

1. Input values are measured at 2D points.

2. Output is a surface representation, raster, or denser/smoothed TIN.

1. Scope of the lecture on spatial interpolation and approximation

1. Caption evidence
   Observation

   Slide title reads “Scattered points to raster” with labels “Given Points” and “Regular Grid”.

2. Audio
   Observation

   The lecturer says they go from scattered or irregularly distributed points to a regular grid, where red points are given points and blue points are interpolated points.

3. Diagram
   Observation

   Left panel shows irregular red points; right panel shows a regular grid of blue points with red given points overlaid.

The example illustrates the basic bivariate interpolation task of transforming an irregular set of known samples into a regular grid. The known samples are visually marked in red, while the interpolated grid nodes are shown in blue, making the distinction between observed data and computed values explicit.

1. Known samples are irregularly distributed.

2. Target output is a regular grid.

1. Definition and applications of bivariate interpolation

1. Caption evidence
   Observation

   Slide text reads “Trivariate: values measured at 3D points to 3D raster” and “Soils properties”.

2. Audio
   Observation

   The lecturer says interpolation can convert values measured at three-dimensional points to three-dimensional raster, using soil properties as an example with missing horizontal points and scattered vertical distribution.

3. Diagram
   Observation

   Left figure shows colored sample points on a 3D grid with an elevation legend; right figures show layered surfaces and a volume labeled “Volume with high %clay”.

Trivariate interpolation extends the same idea from 2D surfaces to 3D volumes: values measured at 3D points are converted into a 3D raster. The soil-properties example emphasizes that sampling may be only partly regular horizontally but scattered vertically, so interpolation is needed to build a continuous volume model from discrete samples.

1. Input values are measured at 3D points.

2. Output is a 3D raster or volume model.

1. Definition and applications of bivariate interpolation

1. Caption evidence
   Observation

   Slide text reads “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)” and “Groundwater pollution over 10 years”.

2. Audio
   Observation

   The lecturer begins discussing even higher dimension interpolation when points exist in three dimensions over time.

3. Diagram
   Observation

   Left figure shows surface points in time with concentration legend and scale bars; right figure shows animated points in 3D space.

1. The spoken explanation is truncated before the lecturer finishes the quadrivariate discussion.

Quadrivariate interpolation is introduced as interpolation from values measured at 3D points over time to a 4D raster that combines three spatial dimensions with time. The groundwater-pollution example over 10 years shows that the method applies to spatiotemporal fields rather than only static spatial surfaces or volumes.

1. Input values are measured at 3D points over time.

2. Output is a 4D raster combining 3D space and time.

1. Trivariate interpolation for 3D soil-property modeling

1. Caption evidence
   Observation

   Slide title “When to use interpolation” and text “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)”.

2. Audio
   Observation

   The lecturer says that if values were measured over time in three-dimensional space, quadrivariate interpolation is needed because there are four variables.

Quadrivariate interpolation is introduced as the method needed when observations are scattered in three-dimensional space and also vary over time. The video explicitly identifies the four variables as 3D space plus time, and the desired output is a 4D raster.

1. Input values are measured at 3D points over time.

2. Output is a 4D raster consisting of 3D space plus time.

1. Caption evidence
   Observation

   Slide text: “Problem definition: find a function which passes through (or close to) the given set of discrete points and then use this function to compute the values at a different set of points – usually a regular grid”.

2. Audio
   Observation

   The lecturer restates that the task is to find a function passing through or close to the given discrete points and then use it to compute values at a different set of points, usually a regular grid.

The video defines interpolation at the level of the whole problem: first construct a function that matches or closely fits a given set of discrete observations, then evaluate that function at another set of locations, typically a regular grid. This separates the fitting stage from the prediction stage.

1. There is a given set of discrete points.

2. A function must pass through or close to those points.

3. The function is then evaluated at a different set of points, usually a regular grid.

1. Formula
   Observation

   Slide text: “Given m-points ($x_i$, $y_i$, $z_i$), $i=1$,m find such $F(x,y)$ that for each $i=1$,m” followed by “$z_i=F(x_i,y_i)$ - interpolation” and “$z_i=F(x_i,y_i) + e(x_i,y_i)$ - approximation”.

2. Audio
   Observation

   The lecturer says they will look at the bivariate formulation for simplicity and describes finding a function that passes through the points.

1. The spoken explanation reaches only the start of the interpolation condition before the clip ends; the full verbal elaboration of both formulas is not present in the audio.

The bivariate formulation takes m observed points ($x_i$, $y_i$, $z_i$) and seeks a function $F(x,y)$. If the function satisfies $z_i = F(x_i,y_i)$ exactly at every sample point, the video labels this interpolation. If instead $z_i = F(x_i,y_i) + e(x_i,y_i)$, with an added error term, the video labels this approximation. The same function is then used to compute $z_k = F(x_k,y_k)$ at grid points ($x_k$,$y_k$).

1. Given m points ($x_i$,$y_i$,$z_i$) with $i=1$,m.

2. Find a function $F(x,y)$.

3. Interpolation requires exact matching at the sample points.

4. Approximation allows an error term $e(x_i,y_i)$.

5. Grid values are computed by evaluating F at ($x_k$,$y_k$).

1. General interpolation problem definition

1. Formula
   Observation

   Slide heading: Problem definition.

2. Formula
   Observation

   Slide text: find a function which passes through (or close to) the given set of discrete points; use this function to compute the values at a different set of points – usually a regular grid.

3. Audio
   Observation

   The lecturer explains finding a function and then using it to compute new z values at grid points.

The video defines the interpolation problem as finding a function that passes through, or close to, a given set of discrete points, and then using that function to compute values at another set of points, usually a regular grid.

1. Input is a discrete set of points.

2. Output is evaluated at a different set of points, usually a regular grid.

1. Formula
   Observation

   Slide heading: Bivariate formulation.

2. Formula
   Observation

   Given m-points ($x_i$, $y_i$, $z_i$), $i=1$,m find such $F(x,y)$ that for each $i=1$,m ...

In the bivariate setting, one is given m points ($x_i$, $y_i$, $z_i$) and seeks a function $F(x,y)$ satisfying the chosen fitting condition at each data location.

1. There are m given data points.

2. The unknown is a function of two variables x and y.

1. Problem definition of interpolation

1. Formula
   Observation

   Slide line labeled interpolation: $z_i=F(x_i,y_i)$.

2. Audio
   Observation

   The lecturer says the value of this function will be the measured value $z_i$.

Interpolation requires the fitted function to match the measured value exactly at every given data point.

1. Applies when exact passage through all data points is required.

1. Bivariate formulation of the interpolation problem

1. Formula
   Observation

   Slide line labeled approximation: $z_i=F(x_i,y_i)+e(x_i,y_i)$.

2. Audio
   Observation

   The lecturer says we can allow a small deviation from this value $z_i$, and this deviation will be this e.

3. Audio
   Observation

   The lecturer adds that approximation is needed if the data are noisy and we are trying to smooth it out.

Approximation relaxes exact fitting by allowing a deviation $e(x_i, y_i)$ between the measured value and the fitted function value. The lecturer explicitly links this relaxation to noisy data and smoothing.

1. Used when the fitted surface need not pass exactly through the data.

2. Motivated in the video by noisy data and smoothing.

1. Bivariate formulation of the interpolation problem

1. Caption evidence
   Observation

   Slide text equates quadrivariate with “3D space + time” and a “4D raster”.

2. Audio
   Observation

   The lecturer says, “We have four variables.”

When data are measured at 3D points over time, the interpolation problem is quadrivariate because it involves four variables: three spatial dimensions and time.

1. Values are measured at 3D points.

2. Measurements occur over time.

For the groundwater example and the stated class of 3D-plus-time data.

1. Formula
   Observation

   The slide explicitly labels “$z_i=F(x_i,y_i)$” as interpolation and “$z_i=F(x_i,y_i)+e(x_i,y_i)$” as approximation.

In the bivariate formulation, interpolation means the fitted function matches the observed values exactly at the sample points, while approximation means the observed values differ from the function by an error term.

1. A function $F(x,y)$ is being fitted to m points ($x_i$,$y_i$,$z_i$).

For each $i=1$,m.

1. Formula
   Observation

   Slide text: Problem does not have a unique solution:

2. Audio
   Observation

   The lecturer says the difficulty is there is no unique solution to this problem.

3. Animation
   Observation

   Several different curves are drawn through the same set of blue points.

Given only the requirement that a function pass through the data points, the interpolation problem does not have a unique solution.

1. A finite set of data points is given.

2. No extra constraints beyond passage through the points are imposed.

There exist multiple functions satisfying the same interpolation condition.

1. Formula
   Observation

   Slide text: we need additional conditions.

2. Audio
   Observation

   The lecturer says to find a unique interpolation function, we need additional conditions.

To obtain a unique interpolation function, additional conditions must be imposed beyond merely passing through the data points.

1. The underlying interpolation problem is non-unique without further restrictions.

For the class of problems discussed in the lecture.

1. Audio
   Observation

   The lecturer says we need this approximation if the data are noisy and we are trying to smooth it out.

2. Formula
   Observation

   The approximation equation includes the deviation term $e(x_i, y_i)$.

Approximation is needed when the data are noisy and the goal is to smooth them out rather than force exact passage through every point.

1. The data contain noise.

2. Smoothing is desired.

Stated as a practical motivation in the lecture.

1. Audio
   Observation

   The lecturer explains that the groundwater wells are scattered in horizontal and vertical directions and also over time, with some wells having only two temporal points and others four or five.

2. Diagram
   Observation

   Two visualizations show points in 3D space and surface points in time for groundwater pollution over 10 years.

1. Expression

   $$$$
   Explanation

   The example data are well measurements located in three-dimensional space.

   Justification

   The slide labels the case as values measured at 3D points and the right visualization is titled “points in3D space”.

   Shown in the video
2. Expression

   $$$$
   Explanation

   The same wells are also sampled repeatedly over time, not just once.

   Justification

   The lecturer states that some wells have only two points over time and others have four or five points over time.

   Shown in the video
3. Expression

   $$$$
   Explanation

   Therefore the data are scattered both in space and in time.

   Justification

   The lecturer explicitly says the data are scattered both in space and time.

   Shown in the video
4. Expression

   $$$$
   Explanation

   To obtain an actual four-dimensional model, one must interpolate in 3D space and time.

   Justification

   This is the lecturer’s concluding statement for the example before moving to the animation.

   Shown in the video

The groundwater pollution example motivates quadrivariate interpolation because the observations vary across three spatial dimensions and time simultaneously.

1. Formula
   Observation

   Slide shows the sequence from given m-points to finding $F(x,y)$, then computing $z_k=F(x_k,y_k)$ where ($x_k$,$y_k$) are grid points.

1. The clip ends before the lecturer verbally develops the “Why is this a difficult problem?” prompt.

1. Expression

   $$(x_i, y_i, z_i),\; i=1,m$$
   Explanation

   Start with m discrete observed points, each having two independent coordinates and one observed value.

   Justification

   This is the given data statement on the bivariate formulation slide.

   Shown in the video
2. Expression

   $$\text{find } F(x,y)$$
   Explanation

   Seek a function of two variables that represents the sample data.

   Justification

   The slide explicitly asks to find such $F(x,y)$.

   Shown in the video
3. Expression

   $$z_i = F(x_i,y_i)$$
   Explanation

   If the function reproduces every observed $z_i$ exactly at its own sample location, the relation is called interpolation.

   Justification

   The slide labels this equation “interpolation”.

   Shown in the video
4. Expression

   $$z_i = F(x_i,y_i) + e(x_i,y_i)$$
   Explanation

   If the observed value differs from the function value by an error term, the relation is called approximation.

   Justification

   The slide labels this equation “approximation”.

   Shown in the video
5. Expression

   $$z_k = F(x_k,y_k)$$
   Explanation

   Once F is chosen, evaluate it at new locations ($x_k$,$y_k$) to produce values on a regular grid.

   Justification

   The slide states “and compute $z_k=F(x_k,y_k)$ where ($x_k$,$y_k$) are grid points”.

   Shown in the video

The bivariate formulation turns discrete samples into predicted grid values by first constructing $F(x,y)$, either exactly interpolating or approximately fitting the data, and then evaluating F at grid locations.

1. Animation
   Observation

   A step-like curve through the blue points is shown first.

2. Animation
   Observation

   A piecewise-linear red curve through the same points is added.

3. Animation
   Observation

   A smooth green curve through the same points is added.

4. Animation
   Observation

   A thick black curve through the same points is added.

5. Audio
   Observation

   The lecturer contrasts steps, spikes, smooth behavior, and wild geometry while noting all functions pass through the blue points.

1. Expression

   $$$$
   Explanation

   Start with a fixed set of blue data points.

   Justification

   These are the given points that any candidate interpolant must pass through.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Draw one admissible function as a step-like curve through the points.

   Justification

   The lecturer presents this as a function that passes through the data points.

   Shown in the video
3. Expression

   $$$$
   Explanation

   Draw a second admissible function as a piecewise-linear curve through the same points.

   Justification

   This shows a different geometric form satisfying the same passage condition.

   Shown in the video
4. Expression

   $$$$
   Explanation

   Draw a third admissible function as a smooth curve through the same points.

   Justification

   The lecturer uses it to contrast with the previous sharper shapes.

   Shown in the video
5. Expression

   $$$$
   Explanation

   Draw a fourth admissible function as a thick black curve through the same points.

   Justification

   The lecturer explicitly says this black function again fulfills the given condition.

   Shown in the video
6. Expression

   $$$$
   Explanation

   Conclude that multiple different functions satisfy the same interpolation condition.

   Justification

   Since all displayed curves pass through the same blue points, the solution cannot be unique without extra constraints.

   Derived from the video

The example sequence proves non-uniqueness constructively by exhibiting several different functions that all interpolate the same data.

1. Audio
   Observation

   The lecturer asks why this is a difficult problem and says finding a function which passes through the data points should be fairly simple.

2. Audio
   Observation

   The lecturer then says the difficulty is there is no unique solution to this problem.

3. Formula
   Observation

   The later slide states: Problem does not have a unique solution: we need additional conditions.

1. Expression

   $$$$
   Explanation

   The lecturer first frames the task as finding a function through the data points.

   Justification

   This matches the earlier problem definition on the slide.

   Shown in the video
2. Expression

   $$$$
   Explanation

   She notes that this part sounds fairly simple.

   Justification

   This is stated directly in the audio.

   Shown in the video
3. Expression

   $$$$
   Explanation

   She identifies the actual difficulty as non-uniqueness of the solution.

   Justification

   The audio explicitly says the difficulty is there is no unique solution.

   Shown in the video
4. Expression

   $$$$
   Explanation

   Therefore the problem becomes difficult not because no interpolant exists, but because many do.

   Justification

   This follows from the visual demonstration of multiple admissible curves.

   Derived from the video

The central difficulty of interpolation is selecting among many valid functions, not merely constructing one.

1. Caption evidence
   Observation

   Slide title “Scattered points to raster”; labels “Given Points” and “Regular Grid”.

2. Audio
   Observation

   The lecturer explicitly identifies red points as given points and blue points as interpolated points.

3. Diagram
   Observation

   Two side-by-side panels compare irregular red samples with a regular blue grid containing the same red samples.

Given an irregular set of measured points, produce values on a regular grid.

1. Red points represent given measurements.

2. The target layout is a regular grid.

Transform scattered observations into interpolated values at regular grid nodes.

1. Expression

   $$$$
   Explanation

   Start from the left panel showing irregularly distributed known points.

   Justification

   The slide labels this panel “Given Points”.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Move to the right panel where a regular grid of blue points is overlaid with the original red points.

   Justification

   The lecturer states that the blue points are interpolated points and the red points are given points.

   Shown in the video

The example demonstrates bivariate interpolation as conversion from scattered 2D samples to a regular raster grid.

Visual comparison of the two panels confirms that the original red samples remain present while additional blue grid nodes are introduced.

1. Caption evidence
   Observation

   Slide text “Trivariate: values measured at 3D points to 3D raster” and “Soils properties”.

2. Audio
   Observation

   The lecturer describes missing horizontal points and scattered vertical distribution, then says a volume model can be created and isolines or cross sections extracted.

3. Diagram
   Observation

   Left figure shows 3D sample points with elevation legend; right figures show layered surfaces and a volume labeled “Volume with high %clay”.

Use 3D point measurements of soil properties to create a continuous 3D model.

1. Measurements are located at 3D points.

2. Horizontal coverage has missing points.

3. Vertical sampling is scattered.

Generate a 3D raster volume from which isolines or cross sections can be extracted.

1. Expression

   $$$$
   Explanation

   Begin with discrete soil-horizon samples plotted in 3D space.

   Justification

   The left figure shows colored sample points associated with elevation.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Apply trivariate interpolation to convert those 3D point values into a 3D raster.

   Justification

   The slide text explicitly states “values measured at 3D points to 3D raster”.

   Shown in the video
3. Expression

   $$$$
   Explanation

   From the resulting volume model, extract derived representations such as layered surfaces, isolines, or cross sections.

   Justification

   The lecturer says these outputs can be extracted from the 3D raster.

   Shown in the video

The example shows trivariate interpolation producing a 3D volume model of soil properties from scattered 3D samples.

The right-side figures display layered surfaces and a volume labeled “Volume with high %clay”, consistent with the stated extraction from a 3D raster.

1. Caption evidence
   Observation

   Slide text “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)” and “Groundwater pollution over 10 years”.

2. Diagram
   Observation

   Left figure shows “surface points in time” with concentration legend and scale bars; right figure shows animated “points in 3D space”.

3. Audio
   Observation

   The lecturer begins introducing higher-dimension interpolation for points in three dimensions over time, but the sentence is cut off.

1. The spoken explanation is incomplete because the clip ends during the introduction.

2. No explicit final result is stated within this fragment.

Represent groundwater pollution measured at 3D points over a 10-year period as a 4D field.

1. Measurements occur at 3D points.

2. The example spans time over 10 years.

3. The target representation is a 4D raster combining 3D space and time.

Introduce quadrivariate interpolation for spatiotemporal data.

1. Expression

   $$$$
   Explanation

   Recognize that the data include both spatial location and time.

   Justification

   The slide text explicitly says “3D points over time to 4D raster (3D space + time)”.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Use the groundwater-pollution example to motivate interpolation beyond static 3D volumes.

   Justification

   The slide title and figures present “Groundwater pollution over 10 years”.

   Shown in the video

The fragment presents quadrivariate interpolation as extension of trivariate interpolation by adding the time dimension.

The visual pairing of “surface points in time” and animated “points in 3D space” supports the stated 3D-space-plus-time formulation.

1. Caption evidence
   Observation

   Slide text: “Groundwater pollution over 10 years” and “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)”.

2. Diagram
   Observation

   Left image labeled “surface points in time”; right image labeled “points in3D space”; later slide “Volume change in time” shows an animated 3D volume evolving within a bounding box.

3. Audio
   Observation

   The lecturer describes wells scattered horizontally and vertically, with different numbers of temporal observations per well, and says a 4D model requires interpolation in 3D space and time.

1. The exact pollutant variable represented by the color scale is not fully legible from the available evidence; the clip clearly supports groundwater pollution over time but not a precise chemical name.

Use scattered groundwater measurements over space and time to motivate what kind of interpolation is needed.

1. Groundwater pollution data over 10 years.

2. Measurements come from wells distributed in 3D space.

3. Each well may have multiple temporal observations.

4. Some wells have only two points over time; others have four or five.

Explain why this dataset calls for quadrivariate interpolation and what the resulting model represents.

1. Expression

   $$$$
   Explanation

   Identify the spatial scatter of the wells in horizontal and vertical directions.

   Justification

   The lecturer explicitly points out scatter in both directions using the 3D-space visualization.

   Shown in the video
2. Expression

   $$$$
   Explanation

   Identify the temporal scatter by noting unequal numbers of repeated measurements per well.

   Justification

   The lecturer says some wells have two temporal points and others have four or five.

   Shown in the video
3. Expression

   $$$$
   Explanation

   Conclude that the data live in 3D space plus time, so the interpolation problem has four variables.

   Justification

   This follows directly from the slide definition of quadrivariate interpolation.

   Shown in the video
4. Expression

   $$$$
   Explanation

   Show the desired result as a 4D raster or evolving volume model.

   Justification

   The slide text names a 4D raster, and the next slide animates volume change in time.

   Shown in the video

The groundwater example requires quadrivariate interpolation because the observations are scattered in three spatial dimensions and over time; the desired output is a 4D raster representing the evolving pollution volume.

The conclusion is checked against the slide text “3D space + time” and the animation titled “Volume change in time”, which visualizes the interpolated 4D result.

1. Diagram
   Observation

   Static title slide followed by a static outline slide listing lecture sections.

2. Audio
   Observation

   The lecturer verbally matches the outline topics.

1. Title slide

2. Outline slide

1. Transition from title slide to outline slide at about 6 seconds.

1. Text content remains static while displayed.

These slides establish the lecture topic and sequence of subtopics without showing mathematical animation.

1. Diagram
   Observation

   Slide titled “When to use interpolation” with text defining bivariate interpolation and two images: a blue point cloud and a colored 3D surface.

2. Animation
   Observation

   A green arrow is drawn from the point cloud toward the surface around 90–95 seconds.

1. Blue scattered point cloud

2. Colored 3D surface

3. Green arrow

1. Green arrow is added from the point cloud to the surface near the end of the interval.

1. The underlying slide text and images remain unchanged.

The arrow visually encodes the transformation from discrete 2D-point measurements to a continuous surface representation.

1. Diagram
   Observation

   Two-panel slide labeled “Given Points” and “Regular Grid”.

2. Animation
   Observation

   A green arrow is drawn from the left panel to the right panel around 105–110 seconds.

1. Left panel with red irregular points

2. Right panel with blue regular grid and red points

3. Green arrow

1. Green arrow appears connecting the left panel to the right panel.

1. Point colors and panel labels remain fixed.

The diagram distinguishes observed samples (red) from interpolated grid nodes (blue) and shows the mapping from irregular sampling to regular raster structure.

1. Diagram
   Observation

   Slide titled “When to use interpolation” with text “Trivariate: values measured at 3D points to 3D raster” and “Soils properties”.

2. Diagram
   Observation

   Left figure includes an elevation legend and 3D sample points; right figures show layered surfaces and a volume labeled “Volume with high %clay”.

3. Animation
   Observation

   A green arrow is drawn from the left figure toward the right-side outputs around 155–165 seconds.

1. 3D sample-point figure

2. Elevation legend

3. Layered surface figures

4. Volume with high %clay

5. Green arrow

1. Green arrow is added from the sample-point figure to the derived 3D outputs.

1. Slide text and figure labels remain static.

The visual sequence represents construction of a 3D raster volume from scattered 3D measurements and subsequent extraction of layered or volumetric features.

1. Diagram
   Observation

   Slide titled “When to use interpolation” with text “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)” and “Groundwater pollution over 10 years”.

2. Animation
   Observation

   Right-side figure animates points in 3D space while the left figure remains static.

1. The clip ends before the lecturer completes the explanation.

1. Left static 3D plot labeled “surface points in time”

2. Right animated plot labeled “points in 3D space”

3. Concentration legend

4. Scale bars

1. Right-side point cloud animates through time-like evolution.

1. Slide text and left figure remain fixed.

The animation illustrates that the field varies not only in 3D space but also over time, motivating a 4D raster representation.

1. Diagram
   Observation

   Left panel labeled “surface points in time” shows colored vertical stems over a green surface with a marked “250 days” length; right panel labeled “points in3D space” shows white points and stems inside a dark 3D volume.

2. Audio
   Observation

   The lecturer contrasts the two views, saying one shows distribution in 3D space with exaggerated depth and the other shows distribution in time where z no longer represents physical location.

1. The exact meaning of the color legend on the left panel is not fully readable from the available evidence.

1. Green topographic or surface plane

2. Colored vertical stems and points

3. Dark 3D volume with white points

4. Labels “surface points in time” and “points in3D space”

5. Marked length “250 days”

1. The lecturer shifts attention from the right 3D-space view to the left time view.

2. In the left view, vertical extent is interpreted as time rather than depth.

3. Different wells are described as having different numbers of temporal observations.

1. Both panels represent the same groundwater dataset.

2. The overall topic remains scattered measurements requiring interpolation.

The two visualizations separate the two sources of scatter: spatial distribution in 3D and temporal repetition at the wells. Together they justify a four-variable interpolation problem.

1. Caption evidence
   Observation

   Slide title “Volume change in time” with subtitle “Groundwater pollution over 10 years - 4D raster (3D space + time)”.

2. Animation
   Observation

   A green upper surface and orange-red subsurface volumes appear inside a red bounding box and change shape over time.

3. Audio
   Observation

   The lecturer says, “you have already seen this animation,” linking it to the need to interpolate in 3D space and time.

1. The exact quantitative meaning of the orange-red volume boundaries is not stated in the clip.

1. Red wireframe bounding box

2. Green surface layer

3. Orange and red subsurface plume volumes

1. The subsurface plume volumes expand, contract, and shift position through the animation.

2. The green surface remains as a reference boundary above the subsurface volumes.

1. The bounding box frame stays fixed.

2. The scene is presented as a 4D raster over 10 years.

The animation visualizes the result of interpolating scattered groundwater measurements into a continuous time-evolving 3D volume field.

1. Caption evidence
   Observation

   Slide title “Interpolation: principles” with section “Problem definition” and the sentence about finding a function through or close to discrete points and computing values at a different set of points, usually a regular grid.

2. Animation
   Observation

   A pointer moves under key phrases such as “function”, “discrete points”, and “different set of points” while the lecturer speaks.

1. Title “Interpolation: principles”

2. Section heading “Problem definition”

3. Pointer or cursor highlighting words in the sentence

1. The pointer emphasizes successive phrases in the definition as the lecturer paraphrases them.

1. The slide remains text-only during this interval.

2. The definition stays focused on function construction and evaluation at new points.

This segment converts the motivating examples into a general mathematical problem statement: fit a function to discrete data, then evaluate it elsewhere, usually on a regular grid.

1. Formula
   Observation

   Slide adds “Bivariate formulation” with “Given m-points ($x_i$, $y_i$, $z_i$), $i=1$,m find such $F(x,y)$ that for each $i=1$,m”, then the two equations for interpolation and approximation, then “and compute $z_k=F(x_k,y_k)$ where ($x_k$,$y_k$) are grid points”, plus the red question “Why is this a difficult problem?”.

2. Animation
   Observation

   A pointer circles or underlines parts of the formulas as the lecturer begins explaining them.

1. The lecturer’s spoken explanation of the formulas is incomplete within the clip.

1. Bivariate formulation equations

2. Pointer marks around $z_i=F(x_i,y_i)$

3. Red question text “Why is this a difficult problem?”

1. Attention moves from the general problem statement to explicit symbolic notation.

2. The distinction between interpolation and approximation is highlighted by separate equations.

1. The setting remains bivariate, with two independent variables x and y.

2. The output locations are identified as grid points ($x_k$,$y_k$).

The slide formalizes the earlier verbal definition: discrete samples lead to a function $F(x,y)$, exact fitting is interpolation, error-based fitting is approximation, and grid prediction uses $z_k=F(x_k,y_k)$.

1. Diagram
   Observation

   Static slide titled Interpolation: principles with problem definition, bivariate formulation, interpolation and approximation equations, and grid-point evaluation.

2. Animation
   Observation

   A pointer moves over the equations and the phrase 'usually a regular grid'.

1. Title 'Interpolation: principles'

2. Problem definition text

3. Bivariate formulation text

4. Equation $z_i=F(x_i,y_i)$

5. Equation $z_i=F(x_i,y_i)+e(x_i,y_i)$

6. Grid-point formula $z_k=F(x_k,y_k)$

7. Pointer

1. Pointer highlights the interpolation equation, then the approximation equation, then the grid-point formula.

1. The slide content remains the same throughout this interval.

This visual establishes the formal definitions of interpolation, approximation, and subsequent evaluation on grid points.

1. Animation
   Observation

   A dark blue step-like curve passes through six light-blue circular data points.

2. Audio
   Observation

   The lecturer says we can draw or find a function which looks like this and it passes through the data points.

1. Six light-blue data points

2. Dark blue step-like curve

3. Pointer

1. The step-like curve is introduced as one admissible solution.

1. The data points remain fixed.

This shows that a discontinuous-looking step function can still satisfy the condition of passing through the given points.

1. Animation
   Observation

   A red piecewise-linear curve is overlaid through the same six data points.

2. Audio
   Observation

   The lecturer says we don't like these steps and want a different kind of function.

1. Same six data points

2. Red piecewise-linear curve

3. Previous step-like curve

1. A new red polygonal curve is added through the same points.

1. The data points remain unchanged.

This provides a second geometrically different interpolant, reinforcing non-uniqueness.

1. Audio
   Observation

   The lecturer lists multiple uses of interpolation: converting scattered points, profiles, or isolines to raster; computing denser or smoother TIN; and resampling continuous-data raster to higher resolution.

2. Caption evidence
   Observation

   Slide text includes both raster conversion and TIN/resampling applications.

One might think spatial interpolation only means turning scattered points into a raster.

The video explicitly presents additional uses: creating denser or smoother TINs and resampling existing continuous-data rasters to higher resolution.

1. Audio
   Observation

   The lecturer states that red points are given points and blue points are interpolated points.

2. Diagram
   Observation

   The right panel overlays blue regular-grid nodes with the original red samples.

In the regular-grid figure, all visible points could be mistaken for measured data.

The video distinguishes red given points from blue interpolated points, so only the red points are original measurements.

1. Audio
   Observation

   The lecturer explicitly says that in the time-distribution view the z value does not represent the actual three-dimensional location but rather distribution in time.

2. Diagram
   Observation

   The left visualization is labeled “surface points in time” and uses vertical extent to indicate days.

Because both visualizations use a vertical direction, one may assume the upward axis always means spatial depth.

The video distinguishes two views: in “points in3D space” the vertical coordinate is depth, but in “surface points in time” the vertical extent encodes time, with example lengths such as 250 days and around 500 days.

1. Caption evidence
   Observation

   The problem definition says to find a function through or close to discrete points and then use that function to compute values at a different set of points, usually a regular grid.

2. Audio
   Observation

   The lecturer frames gridding as the usual target set, not the whole definition.

One might think interpolation is only the act of placing values onto a regular grid.

The video defines interpolation as a two-stage problem: first construct a function from discrete observations, then evaluate that function at other points, which are often but not definitionally restricted to a regular grid.

1. Formula
   Observation

   The slide separately labels “$z_i=F(x_i,y_i)$ - interpolation” and “$z_i=F(x_i,y_i) + e(x_i,y_i)$ - approximation”.

One might use “interpolation” loosely for any fitted surface through data.

In the displayed bivariate formulation, interpolation requires exact equality at the sample points, whereas approximation allows an error term $e(x_i,y_i)$.

1. Audio
   Observation

   The lecturer says finding a function which passes through the data points should be fairly simple.

2. Audio
   Observation

   She immediately follows with: The difficulty is there is no unique solution to this problem.

One may think interpolation is difficult because no function can pass through the data points.

The video states that such functions are easy to find; the real difficulty is that many different functions satisfy the same condition, so the solution is not unique.

1. Audio
   Observation

   The lecturer says we would like it to go smoothly around the points, but then comments that it jumps quite a bit around these points and has a pretty wild geometry.

2. Animation
   Observation

   The green curve is smooth yet still visually extreme between data points.

A smooth curve through the points may be taken as automatically satisfactory.

In the video's example, even a smooth curve can have wild geometry between points, so smoothness by itself is presented as only one possible condition, not a complete solution.

1. Audio
   Observation

   After the outline, the lecturer moves to the first substantive topic, “when do we need to use interpolation”.

2. Caption evidence
   Observation

   Outline lists “spatial interpolation: definition, applications”, then the next slide defines bivariate interpolation.

The general lecture topic contains the specific case of bivariate interpolation as its first application-focused subsection.

1. Audio
   Observation

   The lecturer says “So here is the example that we have already used” after describing conversion of scattered points to raster.

2. Diagram
   Observation

   The next slide is titled “Scattered points to raster”.

The scattered-points-to-raster example is a direct application of the bivariate interpolation definition.

1. Audio
   Observation

   The lecturer transitions from 2D surface examples to “values measured at three-dimensional points to three-dimensional raster”.

2. Caption evidence
   Observation

   Slide changes from bivariate surface examples to “Trivariate: values measured at 3D points to 3D raster”.

Trivariate interpolation generalizes the same interpolation idea from 2D point data and surfaces to 3D point data and volumes.

1. Audio
   Observation

   The lecturer says they can also use even higher dimension interpolation if points exist in three dimensions over time.

2. Caption evidence
   Observation

   Slide text changes to “Quadrivariate: values measured at 3D points over time to 4D raster (3D space + time)”.

1. The spoken explanation is cut off before completion.

Quadrivariate interpolation generalizes trivariate interpolation by adding time as an extra dimension to the 3D spatial field.

1. Caption evidence
   Observation

   The same slide presents “Quadrivariate” and “Groundwater pollution over 10 years” together.

2. Audio
   Observation

   The lecturer uses the groundwater wells as the motivating case for needing four-variable interpolation.

The groundwater pollution dataset is used as the concrete application that motivates quadrivariate interpolation.

1. Caption evidence
   Observation

   The lecture moves from “Problem definition” to “Bivariate formulation” on consecutive slides.

2. Audio
   Observation

   The lecturer says they will look at the bivariate formulation for the sake of simplicity.

The bivariate formulation is the simplified two-variable version of the general interpolation problem stated just before it.

1. Formula
   Observation

   The slide places the exact-fit equation and the error-term equation side by side with different labels.

Within the bivariate formulation, exact matching and error-based fitting are contrasted as interpolation versus approximation.

1. Caption evidence
   Observation

   The problem definition mentions computing values at a different set of points, usually a regular grid; the bivariate slide then writes $z_k=F(x_k,y_k)$ where ($x_k$,$y_k$) are grid points.

The symbolic bivariate formulas depend on the earlier verbal definition: first find a function from discrete points, then evaluate it at new grid locations.

1. Formula
   Observation

   The slide first gives the general problem definition and then the bivariate formulation beneath it.

The bivariate formulation is the specific mathematical version of the general interpolation problem definition shown on the same slide.

1. Formula
   Observation

   Within the bivariate formulation, the line $z_i=F(x_i,y_i)$ is labeled interpolation.

The exact interpolation condition is one of the two alternatives specified inside the bivariate formulation.

1. Formula
   Observation

   Within the bivariate formulation, the line $z_i=F(x_i,y_i)+e(x_i,y_i)$ is labeled approximation.

The approximation condition is the second alternative specified inside the bivariate formulation.

1. Formula
   Observation

   The slide places the interpolation and approximation equations one above the other with contrasting labels.

2. Audio
   Observation

   The lecturer contrasts exact matching with allowing a small deviation e.

Interpolation requires exact equality at the data points, whereas approximation allows a deviation term $e(x_i, y_i)$.

1. Caption evidence
   Observation

   Slide explicitly defines bivariate interpolation.

2. Audio
   Observation

   Lecturer explains its uses for surfaces, rasters, and TINs.

1. Definition and applications of bivariate interpolation
2. Converting scattered points to a regular grid

1. Audio
   Observation

   The lecturer states red points are given points and blue points are interpolated points.

2. Diagram
   Observation

   Two-panel slide contrasts “Given Points” and “Regular Grid”.

1. Scattered points to regular grid example
2. Converting scattered points to a regular grid
3. Do not confuse observed samples with interpolated outputs

1. Caption evidence
   Observation

   Slide text contrasts 2D-point surface interpolation with 3D-point to 3D-raster interpolation.

2. Audio
   Observation

   Lecturer explains soil properties sampled in 3D and modeled as a volume.

1. Definition and applications of bivariate interpolation
2. Trivariate interpolation for 3D soil-property modeling
3. s0-cr-bivariate-to-trivariate

1. Caption evidence
   Observation

   Slide text states “3D points over time to 4D raster (3D space + time)”.

2. Audio
   Observation

   Lecturer begins explaining higher-dimension interpolation but is cut off.

1. The spoken explanation is incomplete in this fragment.

1. Quadrivariate interpolation as 3D space plus time
2. s0-cr-trivariate-to-quadrivariate

1. Audio
   Observation

   The lecturer lists conversion to raster, denser/smoother TIN, and resampling continuous rasters.

2. Caption evidence
   Observation

   Slide bullets list these applications.

1. Definition and applications of bivariate interpolation
2. Interpolation is not only point-to-raster conversion

1. Caption evidence
   Observation

   Slide text defines quadrivariate as 3D space + time.

2. Audio
   Observation

   The lecturer says there are four variables.

1. Quadrivariate interpolation
2. Groundwater pollution over 10 years as a quadrivariate example

1. Audio
   Observation

   The lecturer says the z value in the left view represents distribution in time, not actual 3D location.

2. Diagram
   Observation

   The left visualization is labeled “surface points in time” and includes “250 days”.

1. Quadrivariate interpolation
2. Groundwater pollution over 10 years as a quadrivariate example

1. Caption evidence
   Observation

   Problem definition slide text is fully visible.

2. Audio
   Observation

   The lecturer paraphrases the same definition.

1. General interpolation problem definition

1. Formula
   Observation

   The slide explicitly labels the two equations as interpolation and approximation.

1. Bivariate interpolation and approximation formulation
2. Exact fit versus error-term fit

1. Formula
   Observation

   Slide states “compute $z_k=F(x_k,y_k)$ where ($x_k$,$y_k$) are grid points”.

1. Bivariate interpolation and approximation formulation

1. Formula
   Observation

   Slide heading Problem definition and the text about finding a function through discrete points.

1. Problem definition of interpolation
2. Bivariate formulation of the interpolation problem

1. Formula
   Observation

   Two adjacent equations labeled interpolation and approximation.

2. Audio
   Observation

   The lecturer contrasts exact matching with allowing a small deviation e.

1. Exact interpolation condition
2. Approximation condition with deviation term