DOPPLER-RADAR ESTIMATES OF PRECIPITATION
tinuous variables plotted against one another using RStudio’s “ggplot” (
https://www.rstudio.com). Applying a best-fit line to the data in Figure 2 aids in showing nonlinearity of the dataset. Figure 3 displays the normalized DREP and GEP variables, and bins them by the DREP variable into 36 bands of 0.25 standard deviation increments. Figure 3 is instrumental in setting up an algorithm that examines each band utilizing various statistical analysis methods.
Figure 1 - Raw Doppler Radar (DREP) and Gauge Precipitation (GEP) values collected, arranged, and documented in Table 1.
The Python “Seaborn” (
https://seaborn.pydata.org/) pack- age wrapped around “matplotlib” also performed relevant statistical analysis in this study. The “joint plot” function in Seaborn creates a multi-panel figure that shows both the bivariate (or joint) relationship between the two variables. Figure 4 on page 24 shows the spectrum of the normalized GEP and DREP together with the probability density plot while Figure 5 on page 24 takes a DREP slice, preset to be a 0.25 standard deviation incremental band, and applies a ker- nel density estimate, meaning a nonparametric or unspecific distributed way to estimate the probability density function of the target random variable. This cross-section of the data with respect to the independent variable (DREP), given in terms of standard deviation units, yields the outcome of a frequency-distribution of dependent variable precipitation to be determined.
Figure 2 - Spectrum of the normalized Doppler Radar and Gauge
Precipitation values with a best fit line in red and uncertainty in surrounding dark grey.
The two software programs each provide a high-level of interface for communicating informative statistical graphics. After normalizing the two variables, use of RStudio and Python was central to the analysis of the data. The “joint plot” option in Seaborn aids in analysis and visualization of singular bands of Radar ranges while RStudio’s “ggplot” takes this idea of DREP bands and creates a set of “ridgeline” plots for further analysis and visualization of the entire dataset, consisting of 36 bands of Doppler Radar ranges as seen in Figures 6 and 7 on page 25. In order to accomplish the graphics in these later figures, the first task required is to go back into the original, tabulated data file and conditionally format a new cell that relates to a data pair and creates a responsive bin. The output of this bin categorizes each data point and allows for future plotting ease by preformatting the previously designated 0.25 band increments. In the excel data file: the final “bin” variables follow the label- ing system “Doppler Radar ADJ Band” and “Gauge PPT ADJ Band” relating contained therein. The ggplot function then treats these new “bin” variables as factors and forms the labels for the plotting algorithm to output the useful information as seen previously in the standalone band of Figure 5.
Results
Figure 3 - Spectrum of the normalized Doppler Radar and Gauge Precipitation values binned by increments of 0.25.
www.aipg.org
Treating the Doppler Radar Estimated Precipitation (DREP) variable as the input and the Gauge Estimated Precipitation (GEP) variable as the output, the visualizing of these two variables presents an inverse function when arranging DREP on the vertical axis and GEP on the horizontal axis. This allows for the bands of DREP to present relevant statistical information in the form of kernel density estimates,
Jan.Feb.Mar 2019 • TPG 23
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44 |
Page 45 |
Page 46 |
Page 47 |
Page 48 |
Page 49 |
Page 50 |
Page 51 |
Page 52 |
Page 53 |
Page 54 |
Page 55 |
Page 56 |
Page 57 |
Page 58 |
Page 59 |
Page 60 |
Page 61 |
Page 62 |
Page 63 |
Page 64
