image sizes, links, image aligns

never say last

062-mapping-RK.Rmd

@@ -82,23 +82,23 @@ The purpose of a regression analysis, of course, is to develop a model that can
 **Requirements**
   The following are required to implement Regression Kriging in R
 
-* [Setting-up the Software Environment](Latest version of R software, network connection and sufficient RAM,  storage capacity)
+* Setting-up the Software Environment
 
-* [Obtaining and Installing R Studio](Latest version of RStudio)
+* Obtaining and Installing R Studio
 
-* [R Packages](R packages)
+* R Packages
 
-* [Preparation of local soil property data](Point Dataset)
+* Preparation of local soil property data
 
-* [Preparation of spatial covariates](Environmental predictors)
+* Preparation of spatial covariates
 
-+ [DEM-derived covariates](Relief (e.g. DEM, Slope, TWI))
+  + DEM-derived covariates
 
-+ [Land cover/Land use](Organism map (e.g. land use, NDVI, land cover))
+  + Land cover/Land use
 
-+ [Climate](Climate Data (e.g. mean precipitation, mean temperature))
+  + Climate
 
-+ [Parent material](Parent material (parent material, geology))
+  + Parent material
 
 
 #### Setting Working Space and Initial Steps

07-Validation.Rmd

@@ -236,7 +236,7 @@ We should warn here that the calculation of the $CI$ is based on the assumption
 **Estimation of qualitative map quality measures:** For validation of qualitative soil maps, a sample error matrix is constructed from the validation data (Figure \@ref(fig:errormatrix)). $n$ is the total number of validation locations in the sample. Element $n_{ij}$ of the matrix corresponds to the number of validation locations that have been predicted as class $i$, $i = 1, 2, \dots U$ and belong to class $j$, $j = 1, 2, \dots U$ [@lark1995components]. The matrix summarizes correct predictions and incorrect predictions within the validation data.
 
 
-```{r errormatrix, fig.cap="Sample error matrix" , out.width='80%', echo=FALSE}
+```{r errormatrix, fig.cap="Sample error matrix" , out.width='80%', echo=FALSE, fig.align='center'}
 knitr::include_graphics("images/Validation_error_matrix.png")
 ```
 
@@ -477,19 +477,19 @@ abline(lm(dat$OCSKGM ~ dat$MKD_OCSKGM_svm), col = 'blue', lty=2)
 par(mfrow=c(1,1))
 ```
 
-```{r, fig.cap='Spatial bubble of the prediction errors for RK'}
+```{r, fig.cap='Spatial bubble of the prediction errors for RK', out.width='60%'}
 # spatial bubbles for prediction errors
 bubble(dat[!is.na(dat$PE_RK),], "PE_RK", pch = 21, 
        col=c('red', 'green'))
 ```
 
-```{r, fig.cap='Spatial bubble of the prediction errors for rf'}
+```{r, fig.cap='Spatial bubble of the prediction errors for rf', out.width='60%'}
 # spatial bubbles for prediction errors
 bubble(dat[!is.na(dat$PE_rf),], "PE_rf", pch = 21, 
        col=c('red', 'green'))
 ```
 
-```{r, fig.cap='Spatial bubble of the prediction errors for svm'}
+```{r, fig.cap='Spatial bubble of the prediction errors for svm', out.width='60%'}
 # spatial bubbles for prediction errors
 bubble(dat[!is.na(dat$PE_svm),], "PE_svm", pch = 21, 
        col=c('red', 'green'))
@@ -517,7 +517,7 @@ train.ind <- createDataPartition(1:nrow(dat), p = .75, list = FALSE)
 train <- dat[ train.ind,]
 test  <- dat[-train.ind,]
 
-plot(density (log(train$OCSKGM)), col='red')
+plot(density (log(train$OCSKGM)), col='red', main="")
 lines(density(log(test$OCSKGM)), col='blue')
 legend('topright', legend=c("train", "test"),
       col=c("red", "blue"), lty=1, cex=1.5)

08-model-evaluation.Rmd

@@ -9,7 +9,7 @@ We found in this package a very useful set of functions for model evaluation met
 
 We will import the predicted maps and harmonize them in to the same regular grid (~1x1km of spatial resolution). Then we will plot the statistical distribution and the correlation between the three different methods (RK, RF, SVM).   
 
-```{r}
+```{r, out.width='70%'}
 library(raster)
 RF<-raster('results/MKD_OCSKGM_rf.tif')
 RK<-raster('results/MKD_OCSKGM_RK.tif')

09-uncertainty.Rmd

@@ -61,7 +61,7 @@ In the example above we had chosen the normal distribution because it is the mos
 There are many different soil properties that in addition vary in space and possibly time. Thus, the characterization of uncertainty about soil properties needs to be extended and include cross- and space-time correlations. It is beyond the scope of this chapter to explain this in detail, for this we refer to standard textbooks such as \cite{goovaerts1997geostatistics} and \cite{webster_2007}. If we assume a joint normal distribution, then a vector of soil properties (be it different soil properties or the same soil property at multiple locations, depths or times) $Z$ is fully characterized by the vector of means $m$ and variance-covariance matrix $C$. Figure \@ref(fig:pairedsoils) shows three examples of 500 paired soil property values that were simulated from different bivariate normal distributions. The left panel shows an uncorrelated case with equal standard deviations for both properties. The centre and right panels show a case where soil property 2 has a greater standard deviation than soil property 1. The difference between these two cases is that the centre panel has a zero correlation between the two soil properties while it is positive in the right panel.
 
 
-```{r pairedsoils, fig.cap="Scatter plots of 500 paired soil property values drawn from a two-dimensional normal distribution" , out.width='80%', echo=FALSE}
+```{r pairedsoils, fig.cap="Scatter plots of 500 paired soil property values drawn from a two-dimensional normal distribution" , out.width='80%', echo=FALSE, fig.align='center'}
 knitr::include_graphics("images/pairedsoilpropierties.png")
 ```
 

docs/SOCMapping.tex

@@ -249,8 +249,9 @@ \chapter*{Copyright and disclaimer}\label{copyright-and-disclaimer}}
 www.fao.org/contact-us/licence-request or addressed to
 \href{mailto:copyright@fao.org}{\nolinkurl{copyright@fao.org}}.
 
-FAO information products are available on the FAO website
-(www.fao.org/publications) and can be purchased through
+FAO information products are available on the
+\href{www.fao.org/publications}{FAO website} and can be purchased
+through
 \href{mailto:publications-sales@fao.org}{\nolinkurl{publications-sales@fao.org}}.
 
 \hypertarget{foreword}{%
@@ -3261,28 +3262,26 @@ \subsection{Technical Steps - Regression
 
 \begin{itemize}
 \item
-  \href{Latest\%20version\%20of\%20R\%20software,\%20network\%20connection\%20and\%20sufficient\%20RAM,\%20storage\%20capacity}{Setting-up
-  the Software Environment}
+  Setting-up the Software Environment
 \item
-  \href{Latest\%20version\%20of\%20RStudio}{Obtaining and Installing R
-  Studio}
+  Obtaining and Installing R Studio
 \item
-  \href{R\%20packages}{R Packages}
+  R Packages
 \item
-  \href{Point\%20Dataset}{Preparation of local soil property data}
+  Preparation of local soil property data
 \item
-  \href{Environmental\%20predictors}{Preparation of spatial covariates}
-\item
-  \href{Relief\%20(e.g.\%20DEM,\%20Slope,\%20TWI)}{DEM-derived
-  covariates}
-\item
-  \href{Organism\%20map\%20(e.g.\%20land\%20use,\%20NDVI,\%20land\%20cover)}{Land
-  cover/Land use}
-\item
-  \href{Climate\%20Data\%20(e.g.\%20mean\%20precipitation,\%20mean\%20temperature)}{Climate}
-\item
-  \href{Parent\%20material\%20(parent\%20material,\%20geology)}{Parent
-  material}
+  Preparation of spatial covariates
+
+  \begin{itemize}
+  \item
+    DEM-derived covariates
+  \item
+    Land cover/Land use
+  \item
+    Climate
+  \item
+    Parent material
+  \end{itemize}
 \end{itemize}
 
 \hypertarget{setting-working-space-and-initial-steps}{%
@@ -5886,7 +5885,12 @@ \subsection{Additional Probability
 validation data.
 
 \begin{figure}
-\includegraphics[width=0.8\linewidth]{images/Validation_error_matrix} \caption{Sample error matrix}\label{fig:errormatrix}
+
+{\centering \includegraphics[width=0.8\linewidth]{images/Validation_error_matrix} 
+
+}
+
+\caption{Sample error matrix}\label{fig:errormatrix}
 \end{figure}
 
 From the sample error matrix the overall purity, map unit purity and
@@ -6311,10 +6315,7 @@ \subsection{Graphical Map Quality
 \end{Shaded}
 
 \begin{figure}
-\centering
-\includegraphics{SOCMapping_files/figure-latex/unnamed-chunk-89-1.pdf}
-\caption{\label{fig:unnamed-chunk-89}Spatial bubble of the prediction errors
-for RK}
+\includegraphics[width=0.6\linewidth]{SOCMapping_files/figure-latex/unnamed-chunk-89-1} \caption{Spatial bubble of the prediction errors for RK}\label{fig:unnamed-chunk-89}
 \end{figure}
 
 \begin{Shaded}
@@ -6326,10 +6327,7 @@ \subsection{Graphical Map Quality
 \end{Shaded}
 
 \begin{figure}
-\centering
-\includegraphics{SOCMapping_files/figure-latex/unnamed-chunk-90-1.pdf}
-\caption{\label{fig:unnamed-chunk-90}Spatial bubble of the prediction errors
-for rf}
+\includegraphics[width=0.6\linewidth]{SOCMapping_files/figure-latex/unnamed-chunk-90-1} \caption{Spatial bubble of the prediction errors for rf}\label{fig:unnamed-chunk-90}
 \end{figure}
 
 \begin{Shaded}
@@ -6341,10 +6339,7 @@ \subsection{Graphical Map Quality
 \end{Shaded}
 
 \begin{figure}
-\centering
-\includegraphics{SOCMapping_files/figure-latex/unnamed-chunk-91-1.pdf}
-\caption{\label{fig:unnamed-chunk-91}Spatial bubble of the prediction errors
-for svm}
+\includegraphics[width=0.6\linewidth]{SOCMapping_files/figure-latex/unnamed-chunk-91-1} \caption{Spatial bubble of the prediction errors for svm}\label{fig:unnamed-chunk-91}
 \end{figure}
 
 \hypertarget{dataSplit}{%
@@ -6364,7 +6359,7 @@ \subsection{Data-Splitting}\label{dataSplit}}
 \NormalTok{train <-}\StringTok{ }\NormalTok{dat[ train.ind,]}
 \NormalTok{test  <-}\StringTok{ }\NormalTok{dat[}\OperatorTok{-}\NormalTok{train.ind,]}
 
-\KeywordTok{plot}\NormalTok{(}\KeywordTok{density}\NormalTok{ (}\KeywordTok{log}\NormalTok{(train}\OperatorTok{$}\NormalTok{OCSKGM)), }\DataTypeTok{col=}\StringTok{'red'}\NormalTok{)}
+\KeywordTok{plot}\NormalTok{(}\KeywordTok{density}\NormalTok{ (}\KeywordTok{log}\NormalTok{(train}\OperatorTok{$}\NormalTok{OCSKGM)), }\DataTypeTok{col=}\StringTok{'red'}\NormalTok{, }\DataTypeTok{main=}\StringTok{""}\NormalTok{)}
 \KeywordTok{lines}\NormalTok{(}\KeywordTok{density}\NormalTok{(}\KeywordTok{log}\NormalTok{(test}\OperatorTok{$}\NormalTok{OCSKGM)), }\DataTypeTok{col=}\StringTok{'blue'}\NormalTok{)}
 \KeywordTok{legend}\NormalTok{(}\StringTok{'topright'}\NormalTok{, }\DataTypeTok{legend=}\KeywordTok{c}\NormalTok{(}\StringTok{"train"}\NormalTok{, }\StringTok{"test"}\NormalTok{),}
       \DataTypeTok{col=}\KeywordTok{c}\NormalTok{(}\StringTok{"red"}\NormalTok{, }\StringTok{"blue"}\NormalTok{), }\DataTypeTok{lty=}\DecValTok{1}\NormalTok{, }\DataTypeTok{cex=}\FloatTok{1.5}\NormalTok{)}
@@ -6450,7 +6445,7 @@ \section{Technical steps - Model correlations and spatial
 \end{Highlighting}
 \end{Shaded}
 
-\includegraphics{SOCMapping_files/figure-latex/unnamed-chunk-95-1.pdf}
+\includegraphics[width=0.7\linewidth]{SOCMapping_files/figure-latex/unnamed-chunk-95-1}
 Here we found that the higher correlation between predicted values was
 between RK and SVM (0.86). We also found that the statistical
 distribution of predicted values is quite similar between the three
@@ -7004,7 +6999,12 @@ \subsection{Uncertainty Characterised by Probability
 while it is positive in the right panel.
 
 \begin{figure}
-\includegraphics[width=0.8\linewidth]{images/pairedsoilpropierties} \caption{Scatter plots of 500 paired soil property values drawn from a two-dimensional normal distribution}\label{fig:pairedsoils}
+
+{\centering \includegraphics[width=0.8\linewidth]{images/pairedsoilpropierties} 
+
+}
+
+\caption{Scatter plots of 500 paired soil property values drawn from a two-dimensional normal distribution}\label{fig:pairedsoils}
 \end{figure}
 
 \hypertarget{propagation-of-model-uncertainty}{%

docs/chvalidation.html

@@ -500,7 +500,7 @@ <h3><span class="header-section-number">7.4.1</span> Additional Probability Samp
 \end{equation}\]</span>
 <p>We should warn here that the calculation of the <span class="math inline">\(CI\)</span> is based on the assumption that the estimated map quality measure means have a normal distribution (the central limit theorem). For the squared errors this assumption can be unrealistic, especially for small sample sizes.</p>
 <p><strong>Estimation of qualitative map quality measures:</strong> For validation of qualitative soil maps, a sample error matrix is constructed from the validation data (Figure <a href="chvalidation.html#fig:errormatrix">7.2</a>). <span class="math inline">\(n\)</span> is the total number of validation locations in the sample. Element <span class="math inline">\(n_{ij}\)</span> of the matrix corresponds to the number of validation locations that have been predicted as class <span class="math inline">\(i\)</span>, <span class="math inline">\(i = 1, 2, \dots U\)</span> and belong to class <span class="math inline">\(j\)</span>, <span class="math inline">\(j = 1, 2, \dots U\)</span> <span class="citation">(Lark <a href="#ref-lark1995components">1995</a>)</span>. The matrix summarizes correct predictions and incorrect predictions within the validation data.</p>
-<div class="figure"><span id="fig:errormatrix"></span>
+<div class="figure" style="text-align: center"><span id="fig:errormatrix"></span>
 <img src="images/Validation_error_matrix.png" alt="Sample error matrix" width="80%" />
 <p class="caption">
 Figure 7.2: Sample error matrix
@@ -820,7 +820,7 @@ <h3><span class="header-section-number">7.5.3</span> Graphical Map Quality Measu
 <a class="sourceLine" id="cb159-2" data-line-number="2"><span class="kw">bubble</span>(dat[<span class="op">!</span><span class="kw">is.na</span>(dat<span class="op">$</span>PE_RK),], <span class="st">&quot;PE_RK&quot;</span>, <span class="dt">pch =</span> <span class="dv">21</span>, </a>
 <a class="sourceLine" id="cb159-3" data-line-number="3">       <span class="dt">col=</span><span class="kw">c</span>(<span class="st">&#39;red&#39;</span>, <span class="st">&#39;green&#39;</span>))</a></code></pre></div>
 <div class="figure"><span id="fig:unnamed-chunk-89"></span>
-<img src="SOCMapping_files/figure-html4/unnamed-chunk-89-1.png" alt="Spatial bubble of the prediction errors for RK" width="672" />
+<img src="SOCMapping_files/figure-html4/unnamed-chunk-89-1.png" alt="Spatial bubble of the prediction errors for RK" width="60%" />
 <p class="caption">
 Figure 7.5: Spatial bubble of the prediction errors for RK
 </p>
@@ -829,7 +829,7 @@ <h3><span class="header-section-number">7.5.3</span> Graphical Map Quality Measu
 <a class="sourceLine" id="cb160-2" data-line-number="2"><span class="kw">bubble</span>(dat[<span class="op">!</span><span class="kw">is.na</span>(dat<span class="op">$</span>PE_rf),], <span class="st">&quot;PE_rf&quot;</span>, <span class="dt">pch =</span> <span class="dv">21</span>, </a>
 <a class="sourceLine" id="cb160-3" data-line-number="3">       <span class="dt">col=</span><span class="kw">c</span>(<span class="st">&#39;red&#39;</span>, <span class="st">&#39;green&#39;</span>))</a></code></pre></div>
 <div class="figure"><span id="fig:unnamed-chunk-90"></span>
-<img src="SOCMapping_files/figure-html4/unnamed-chunk-90-1.png" alt="Spatial bubble of the prediction errors for rf" width="672" />
+<img src="SOCMapping_files/figure-html4/unnamed-chunk-90-1.png" alt="Spatial bubble of the prediction errors for rf" width="60%" />
 <p class="caption">
 Figure 7.6: Spatial bubble of the prediction errors for rf
 </p>
@@ -838,7 +838,7 @@ <h3><span class="header-section-number">7.5.3</span> Graphical Map Quality Measu
 <a class="sourceLine" id="cb161-2" data-line-number="2"><span class="kw">bubble</span>(dat[<span class="op">!</span><span class="kw">is.na</span>(dat<span class="op">$</span>PE_svm),], <span class="st">&quot;PE_svm&quot;</span>, <span class="dt">pch =</span> <span class="dv">21</span>, </a>
 <a class="sourceLine" id="cb161-3" data-line-number="3">       <span class="dt">col=</span><span class="kw">c</span>(<span class="st">&#39;red&#39;</span>, <span class="st">&#39;green&#39;</span>))</a></code></pre></div>
 <div class="figure"><span id="fig:unnamed-chunk-91"></span>
-<img src="SOCMapping_files/figure-html4/unnamed-chunk-91-1.png" alt="Spatial bubble of the prediction errors for svm" width="672" />
+<img src="SOCMapping_files/figure-html4/unnamed-chunk-91-1.png" alt="Spatial bubble of the prediction errors for svm" width="60%" />
 <p class="caption">
 Figure 7.7: Spatial bubble of the prediction errors for svm
 </p>
@@ -855,7 +855,7 @@ <h3><span class="header-section-number">7.5.4</span> Data-Splitting</h3>
 <a class="sourceLine" id="cb162-6" data-line-number="6">train &lt;-<span class="st"> </span>dat[ train.ind,]</a>
 <a class="sourceLine" id="cb162-7" data-line-number="7">test  &lt;-<span class="st"> </span>dat[<span class="op">-</span>train.ind,]</a>
 <a class="sourceLine" id="cb162-8" data-line-number="8"></a>
-<a class="sourceLine" id="cb162-9" data-line-number="9"><span class="kw">plot</span>(<span class="kw">density</span> (<span class="kw">log</span>(train<span class="op">$</span>OCSKGM)), <span class="dt">col=</span><span class="st">&#39;red&#39;</span>)</a>
+<a class="sourceLine" id="cb162-9" data-line-number="9"><span class="kw">plot</span>(<span class="kw">density</span> (<span class="kw">log</span>(train<span class="op">$</span>OCSKGM)), <span class="dt">col=</span><span class="st">&#39;red&#39;</span>, <span class="dt">main=</span><span class="st">&quot;&quot;</span>)</a>
 <a class="sourceLine" id="cb162-10" data-line-number="10"><span class="kw">lines</span>(<span class="kw">density</span>(<span class="kw">log</span>(test<span class="op">$</span>OCSKGM)), <span class="dt">col=</span><span class="st">&#39;blue&#39;</span>)</a>
 <a class="sourceLine" id="cb162-11" data-line-number="11"><span class="kw">legend</span>(<span class="st">&#39;topright&#39;</span>, <span class="dt">legend=</span><span class="kw">c</span>(<span class="st">&quot;train&quot;</span>, <span class="st">&quot;test&quot;</span>),</a>
 <a class="sourceLine" id="cb162-12" data-line-number="12">      <span class="dt">col=</span><span class="kw">c</span>(<span class="st">&quot;red&quot;</span>, <span class="st">&quot;blue&quot;</span>), <span class="dt">lty=</span><span class="dv">1</span>, <span class="dt">cex=</span><span class="fl">1.5</span>)</a></code></pre></div>

docs/index.html

@@ -351,7 +351,7 @@ <h1>Copyright and disclaimer</h1>
 <p>FAO, 2018</p>
 <p>FAO encourages the use, reproduction and dissemination of material in this information product. Except where otherwise indicated, material may be copied, downloaded and printed for private study, research and teaching purposes, or for use in non-commercial products or services, provided that appropriate acknowledgement of FAO as the source and copyright holder is given and that FAO’s endorsement of users’ views, products or services is not implied in any way.</p>
 <p>All requests for translation and adaptation rights, and for resale and other commercial use rights should be made via www.fao.org/contact-us/licence-request or addressed to <a href="mailto:copyright@fao.org">copyright@fao.org</a>.</p>
-<p>FAO information products are available on the FAO website (www.fao.org/publications) and can be purchased through <a href="mailto:publications-sales@fao.org">publications-sales@fao.org</a>.</p>
+<p>FAO information products are available on the <a href="www.fao.org/publications">FAO website</a> and can be purchased through <a href="mailto:publications-sales@fao.org">publications-sales@fao.org</a>.</p>
 </div>
             </section>