RMSE Night LST Vs TFA + cfsv2 anomalies iteration

// Change the path to your own TFA images
var MODIS_TFA = ee.Image("users/shilosh/MODIS_LST_TFA_Night"),
    CFSV2_TFA = ee.Image("users/shilosh/CFSv2_LST_TFA_Night");
    
// create list of years over which to iterate
var years = ee.List.sequence(2002, 2019);
// Iterate through years 
var correlation_statistics_by_year = years.map(function(year) {
// var year = 2017
  var calculation_state = 'TFA plus anomalies'
  // var calculation_state = 'TFA only'
  var firstDay = ee.Algorithms.String(ee.Number(year).toInt()).cat('-01-01');
  var lastDay  = ee.Algorithms.String(ee.Number(year).add(1).toInt()).cat('-01-01');
  var geometry = ee.Geometry.Rectangle([33.2,29.0,36.6,34.0]);
  var geometry_json = geometry.toGeoJSON(); 
  
  // Reverse the images into imageCollections
  var MODIS_TFA_ic = ee.ImageCollection(MODIS_TFA.bandNames().map(function(name) { 
    return MODIS_TFA.select([ee.Algorithms.String(name)],['mod']).set('system:DOY', ee.Number.parse(ee.Algorithms.String(name).replace('TFA','0').replace('_','')).add(1)) }))
  var CFSV2_TFA_ic = ee.ImageCollection(CFSV2_TFA.bandNames().map(function(name) { 
    return CFSV2_TFA.select([ee.Algorithms.String(name)],['cfs']).set('system:DOY', ee.Number.parse(ee.Algorithms.String(name).replace('TFA','0').replace('_','')).add(1)) }))
  
  //convert Kelvin to Celsius
  var modis_k2celsius = function(image) {
        return image.multiply(ee.Image(0.02))
                    .subtract(ee.Image(273.15))
                    .set('system:time_start', image.get('system:time_start'));
  };
  
  // Reproject to CFSv2 projection and reduce resolution
  // Get information about the CFSV2 projection.
  var Reproject = function(img){
   var CFSV2Projection = CFSV2_TFA.projection();
   var scale = CFSV2_TFA.projection().nominalScale().getInfo();
  // Get the MODIS data at CFSV2 scale and projection.
  var mod = img
      // Force the next reprojection to aggregate instead of resampling.
      .reduceResolution({
        reducer: ee.Reducer.mean(),
        maxPixels: 10000
      })
      // Request the data at the scale and projection of the CFSV2 image.
      .reproject({
        crs: CFSV2Projection,
        scale: scale 
      });
      return mod;
  };
  
  // Add a property with doy to the colection.
  function createDoyBand(img) {
    var d = ee.Date(img.get('system:time_start'))
      .getRelative('day', 'year')
      .add(1);
    img=img.set('system:DOY', d);
    return img;
  }
  
  // Construct image date from 'system:index' and add it to a new 'date' property 
  var addTimeStampToModis = function(image) {
    var start = ee.String(image.get('system:index'));
    var y = start.slice(0,4);
    var m = start.slice(5,7);
    var d = start.slice(8,10);
    var date = y.cat('-').cat(m).cat('-').cat(d);
    return image.set({'system:time_start': date});
  };
  
  // Construct image date from 'system:index' and add it to a new 'date' property 
  var addTimeStampToCFSv2 = function(image) {
     var start = ee.String(image.get('system:index'));
     var y = start.slice(0,4);
     var m = start.slice(4,6);
     var d = start.slice(6,8);
     var date = y.cat('-').cat(m).cat('-').cat(d);
     return image.set({'system:time_start': date});
  };
  
  // Mask modis
  var mask_good_quality = function(image) {
    var QCNight = image.select('QC_Night');
    var QC = (QCNight.neq(2)).and(QCNight.neq(3));
    return image.updateMask(QC)
                .select('LST_Night_1km')
                .set('system:time_start', image.get('system:time_start'))
  };
  
  // Load imageCollection of MODIS LST
  var MODIS_LST = ee.ImageCollection('MODIS/006/MYD11A1')
              .filterDate(firstDay, lastDay) 
              .select('LST_Night_1km', 'QC_Night')
              .map(mask_good_quality)
              .map(modis_k2celsius)
              // .map(Reproject)
              .map(addTimeStampToModis)
              .map(createDoyBand);
  // var filter_low_values = require('users/shilosh/default:filter_modis_low_values_func');
  // MODIS_LST = filter_low_values.filter_modis_low_values(MODIS_LST, -9);
  
  //convert Kelvin to Celsius
  var k2celsius = function(image) {
        return image.subtract(ee.Image(273.15))   
                    .set('system:time_start', image.get('system:time_start'));
  };
  
  // Load imageCollection of CFSv2 Temperature over ground
  var CFSV2 = ee.ImageCollection('NOAA/CFSV2/FOR6H')
              .filterDate(firstDay, lastDay)
              .select('Maximum_temperature_height_above_ground_6_Hour_Interval')
              .filter(ee.Filter.stringEndsWith('system:index', '00'))
              .map(k2celsius)
              // .map(Reproject)
              .map(addTimeStampToCFSv2)
              .map(createDoyBand);
  
  // Use an equals filter to specify how the collections match.
  var Filter = ee.Filter.equals({
    leftField: 'system:DOY',
    rightField: 'system:DOY'
  });
    // Define the join.
    var innerJoin = ee.Join.inner('primary', 'secondary');
  
  // Join CFSV2 with CFSV2_TFA_ic by DOY
    // Apply the join.
    var CFSV2_JoinInner = innerJoin.apply(CFSV2, CFSV2_TFA_ic, Filter);
    
  // Calculate CFSv2 anomalies
  var CFSV2_Anomalies = CFSV2_JoinInner.map(function(f) {
    var tfa = ee.Image(f.get('secondary'));
    var actual = ee.Image(f.get('primary'));
    return actual.subtract(tfa);
  }).map(addTimeStampToCFSv2)
                 .map(createDoyBand);
  
  // Join MODIS_TFA_ic with CFSV2_Anomalies by DOY
    // Apply the join.
    var MODIS_JoinInner = innerJoin.apply(CFSV2_Anomalies, MODIS_TFA_ic, Filter);
  
  // Calculate MODIS TFA Plus CFSv2 anomalies
  var MODIS_Continuous = MODIS_JoinInner.map(function(f) {
    var tfa       = ee.Image(f.get('secondary'));
    var anomalies = ee.Image(f.get('primary'));
    switch(calculation_state){
      case 'TFA plus anomalies':
        return anomalies.add(tfa);
      case 'TFA only':
        return anomalies.add(tfa).subtract(anomalies);
    }
  }).map(addTimeStampToCFSv2)
                 .map(createDoyBand);
  
  // Join MODIS_Anomalies with CFSV2_Anomalies by DOY
    // Apply the join.
    var Continuous_JoinInner = innerJoin.apply(MODIS_LST, MODIS_Continuous, Filter);

  // Mask CFSv2 
  var MaskedValues = Continuous_JoinInner.map(function(f) {
    var con = ee.Image(f.get('secondary'));
    var mod = ee.Image(f.get('primary'));
    con = ee.Image(con.float()).updateMask(ee.Image(mod).add(10));
    mod = ee.Image(mod.float()).updateMask(ee.Image(con).add(10));
    var mod_vs_cfs = ee.Image(mod).addBands(con);
    mod_vs_cfs = ee.Image(mod_vs_cfs).select(['LST_Night_1km','Maximum_temperature_height_above_ground_6_Hour_Interval'],['mod','con'])
    // mod_vs_cfs = ee.Image(mod_vs_cfs).select(['LST_Day_1km','Temperature_height_above_ground'],['mod','cfs'])
    // mod_vs_cfs = ee.Image(mod_vs_cfs).select(['LST_Day_1km'],['mod'])
    // var image = ee.Image(mod_vs_cfs).select(['LST_Day_1km','Temperature_height_above_ground'],['mod','cfs']).reduceRegion(ee.Reducer.toList());
    // var x = image.get('mod');
    // var y = image.get('cfs');
    return mod_vs_cfs
  });

  var differences_squared = function(image) {
    var differences = ee.Image(image).select('con').subtract(ee.Image(image).select('mod') )
    differences = ee.Image(differences).multiply(differences);
    return differences;
  };
  var mean_of_differences_squared = MaskedValues.map(differences_squared)
  mean_of_differences_squared = ee.ImageCollection(mean_of_differences_squared).mean()                                              
  // print('mean_of_differences_squared = ', mean_of_differences_squared)
  var rmse = mean_of_differences_squared.sqrt()
  
  var mean_absolut_error = function(image) {
    var differences = ee.Image(image).select('con').subtract(ee.Image(image).select('mod') )
    differences = ee.Image(differences).abs();
    return differences;
  };
  var MAE = MaskedValues.map(mean_absolut_error)
  MAE = ee.ImageCollection(MAE).mean()                                              
  
  var Pearson = ee.ImageCollection(MaskedValues).reduce(ee.Reducer.pearsonsCorrelation());
  
  var scale = ee.Image('MODIS/006/MYD11A1/2018_01_01').projection().nominalScale().getInfo();
  var scale = ee.Image('NOAA/CFSV2/FOR6H/2018010112').projection().nominalScale().getInfo();
  var CFSV2Projection = CFSV2_TFA.projection();
  // var scale = CFSV2_TFA.projection().nominalScale().getInfo();
  var rmse_val = rmse.reduceRegion(ee.Reducer.mean(), geometry, scale)
  var MAE_val = MAE.reduceRegion(ee.Reducer.mean(), geometry_json, scale)
  var Pearson_val = Pearson.select('correlation').reduceRegion({
    reducer:ee.Reducer.mean(),
    geometry:geometry_json,
    scale:scale,
    crs:CFSV2Projection,
    })

// print('RMSE = ', rmse_val)  
  var f = ee.Feature(null,{
        'year'    : year,
        'Pearson' : Pearson_val,
        'RMSE'    : rmse_val,
        'MAE'     : MAE_val,
      })
  return f;
})

var fromList = ee.FeatureCollection(correlation_statistics_by_year);
print(fromList);
// Export the FeatureCollection to a CSV file.
Export.table.toDrive({
  collection: fromList,
  // description:'Night_TFA_only',
  description:'Night_TFA_plus_anomalies',
  folder: 'gee',
  fileFormat: 'CSV'
});