Wrong values using gdal_calc.py

Question

I'm using gdal_calc.py to do some raster operations. The raster files stores precipitation values that are between 0 - 40 l/m² aprox. in some cases and the operations i'm trying to do:

1. Values greater than 20 --> reclasify to 20.
2. Values lower than 0 --> reclasify to 0.
3. Average of the reclasified rasters.

Here are some information of the inputs raster:

• RASTER "A":

Band 1 Block=256x16 Type=Float32, ColorInterp=undefined Description = 1072016p

Min=0 Max=0.399 NoData Value=-3.4028234663852886e+38

• RASTER "B"

Band 1 Block=256x16 Type=Float32, ColorInterp=undefined Description = 972016p

Min=0 Max=2.599 NoData Value=-3.4028234663852886e+38

This is what I tried:

gdal_calc.py -A ./datos/10072016p -B ./datos/9072016p  --calc="20*(A>20)" --calc="0*(A<0)" --calc="20*(B>20)" --calc="0*(B<0)" --calc="A+B/2" --outfile=result.tiff

And the result i'm getting:

Band 1 Block=3922x1 Type=Float32, ColorInterp=Gray
Min=-4294967295 Max=4294967295 NoData Value=1.175494351e-38

So, as you can see is not the expected result? What i'm missing or what i'm doing wrong?

EDIT 17/08/2016

After some work trying to understand what's happening I've conclude that the operations that was failling was 0*(A<0), 20*(A>20), ... It works if I just calculate the average of all rasters:

(A + B + C + ...)/length

So the problem, really is that I can't reclassify the values lower than 0 and greater than 20.

Here is the approach i'm using that is working (NodeJS Script) :

const fs = require('fs'),
Exec = require('child_process').exec,
path = require('path'),
dataDir = './datos', // Directorio de donde coger los archivos raster
reg = /.*p$/, // Regex detecto p al final del archivo

// Devuelve una letra entre A-Z y A1-Z1 si se supera length > 26
getNums = length => length <= 26 ? 
    Array.from({length}, (el, i)=> String.fromCharCode(i + 65)) : 
    Array.from({length : 26}, (el, i)=> String.fromCharCode(i + 65)).concat(Array.from({length : length - 26}, (el, i)=> String.fromCharCode(i + 65) + 1)),

// Comprueba que sea el tipo de archivo que se quiere
_RasterFile = file => reg.exec(file),
// Coge el fichero raster y obtiene la fecha
_ParseDate = date => new Date(
    +date.substring(date.length - 5, date.length - 1), 
    +date.substring(date.length - 7, date.length - 5) - 1, 
    +date.substring(0, date.length - 7)
),
// Fecha de hoy
today = new Date(),
// Fecha hace 40 días
last40days = new Date();
last40days.setDate(last40days.getDate() - 40);

// Lista de archivos finales
let listFiles = [];
// nombre del archivo resultante
let resultRasterName = `${today.getDate()}${today.getMonth() + 1 < 10 ? '0' + (today.getMonth() + 1) : today.getMonth() + 1}${today.getFullYear()}rp.tiff`;
fs.readdir(dataDir, (err, files)=>{
  if(err) return console.error(err);
  // Ficheros filtrados con el Regex
  let rasterFiles = files.filter(_RasterFile);
  // Fecha de los ficheros anteriores
  let dateList = rasterFiles.map(_ParseDate);
  // Puesta en común de las dos listas
  listFiles = rasterFiles.map( (f, i) => ({
    name    : f,
    path    : dataDir + '/' + f,
    date    : dateList[i]
  }))
  // Cogiendo los ficheros de los últimos 40 días
  .filter( f => f.date.getTime() >= last40days.getTime() )
  // Ordenándolas por fecha
  .sort( (a, b) => b.date.getTime() - a.date.getTime() );

  // Array de letras en base a el número de archivos
  let ArrayLetras = getNums(listFiles.length);
  // Expresión de los cálculos a realizar ej : (A + B + C + ...)/length
  let calc = '(' + ArrayLetras.join(' + ') + ')/' + ArrayLetras.length;

  // Comando final que se ejecutará ej: gdal_calc.py -A ./datos/10072016p -B ... --calc="expresión" --outfile nombreArchivoFinal.tiff
  let exec = listFiles.reduce( (execString, file, idx) => execString + '-' + ArrayLetras[idx] + ' ' + file.path + ' ', 'gdal_calc.py ') + 
    '--calc="' + calc + '" --outfile ' + resultRasterName;

  console.log('executing...', exec);

  Exec(exec, (error, stdout, stderr)=>{
    console.log(stdout);
  });
})

Answer 1

If you use a simple --calc option gdal_calc.py should work fine:

gdal_calc.py -A ./datos/10072016p -B ./datos/9072016p --calc="(20*(A>20)+0*(A<0)+20*(B>20)+0*(B<0))/2" --outfile=result.tiff