Read temperature and salinity from the PML POLCOMS-ERSEM NetCDF model
output files and interpolate onto the open boundaries in Mobj.
function Mobj = get_POLCOMS_tsobc(Mobj, ts, polcoms_bathy, varlist)
DESCRIPTION:
Interpolate temperature and salinity values onto the FVCOM open
boundaries at all sigma levels.
INPUT:
Mobj = MATLAB mesh structure which must contain:
- Mobj.siglayz - sigma layer depths for all model nodes.
- Mobj.lon, Mobj.lat - node coordinates (lat/long).
- Mobj.obc_nodes - list of open boundary node inidices.
- Mobj.nObcNodes - number of nodes in each open boundary.
ts = Cell array of PML POLCOMS-ERSEM NetCDF file(s) in which 4D
variables of temperature and salinity (called 'ETWD' and
'x1XD') exist. Their shape should be (y, x, sigma, time).
NOTES:
- If you supply multiple files in ts, there are a few assumptions:
- Variables are only appended if there are 4 dimensions; fewer than
that, and the values are assumed to be static across all the given
files (e.g. longitude, latitude). The fourth dimension is time.
- The order of the files given should be chronological.
- The NetCDF files used here are those from the PML POLCOMS-ERSEM model
output.
OUTPUT:
Mobj = MATLAB structure in which three new fields (called temperature,
salinity and ts_times). temperature and salinity have sizes
(sum(Mobj.nObcNodes), sigma, time). The time dimension is
determined based on the input NetCDF file. The ts_time variable
is just the input file times in Modified Julian Day.
EXAMPLE USAGE
Mobj = get_POLCOMS_tsobc(Mobj, ts)
Author(s):
Pierre Cazenave (Plymouth Marine Laboratory)
Revision history
2013-02-07 First version.
2013-02-27 Change the vertical interpolation to be scaled within the
POLCOMS-ERSEM depth range for the current node. The net result is that
the vertical profiles are squashed or stretched to fit within the
FVCOM depths. This means the full profile structure is maintained in
the resulting FVCOM boundary input despite the differing depths at the
FVCOM boundary node.
2013-06-03 Fix some bugs in the way the open boundary node values were
stored (the order in which they were stored did not match the order of
the nodes in Casename_obc.dat). Also fix the order of the vertically
interpolated values so that FVCOM starts at the surface instead of
mirroring POLCOMS' approach (where the first value is the seabed). The
effect of these two fixes (nodes and vertical) should match what FVCOM
expects. Also add a set of figures (commented out) at the end for
diagnostic purposes.
==========================================================================0001 function Mobj = get_POLCOMS_tsobc(Mobj, ts) 0002 % Read temperature and salinity from the PML POLCOMS-ERSEM NetCDF model 0003 % output files and interpolate onto the open boundaries in Mobj. 0004 % 0005 % function Mobj = get_POLCOMS_tsobc(Mobj, ts, polcoms_bathy, varlist) 0006 % 0007 % DESCRIPTION: 0008 % Interpolate temperature and salinity values onto the FVCOM open 0009 % boundaries at all sigma levels. 0010 % 0011 % INPUT: 0012 % Mobj = MATLAB mesh structure which must contain: 0013 % - Mobj.siglayz - sigma layer depths for all model nodes. 0014 % - Mobj.lon, Mobj.lat - node coordinates (lat/long). 0015 % - Mobj.obc_nodes - list of open boundary node inidices. 0016 % - Mobj.nObcNodes - number of nodes in each open boundary. 0017 % ts = Cell array of PML POLCOMS-ERSEM NetCDF file(s) in which 4D 0018 % variables of temperature and salinity (called 'ETWD' and 0019 % 'x1XD') exist. Their shape should be (y, x, sigma, time). 0020 % 0021 % NOTES: 0022 % 0023 % - If you supply multiple files in ts, there are a few assumptions: 0024 % 0025 % - Variables are only appended if there are 4 dimensions; fewer than 0026 % that, and the values are assumed to be static across all the given 0027 % files (e.g. longitude, latitude). The fourth dimension is time. 0028 % - The order of the files given should be chronological. 0029 % 0030 % - The NetCDF files used here are those from the PML POLCOMS-ERSEM model 0031 % output. 0032 % 0033 % OUTPUT: 0034 % Mobj = MATLAB structure in which three new fields (called temperature, 0035 % salinity and ts_times). temperature and salinity have sizes 0036 % (sum(Mobj.nObcNodes), sigma, time). The time dimension is 0037 % determined based on the input NetCDF file. The ts_time variable 0038 % is just the input file times in Modified Julian Day. 0039 % 0040 % EXAMPLE USAGE 0041 % Mobj = get_POLCOMS_tsobc(Mobj, ts) 0042 % 0043 % Author(s): 0044 % Pierre Cazenave (Plymouth Marine Laboratory) 0045 % 0046 % Revision history 0047 % 2013-02-07 First version. 0048 % 2013-02-27 Change the vertical interpolation to be scaled within the 0049 % POLCOMS-ERSEM depth range for the current node. The net result is that 0050 % the vertical profiles are squashed or stretched to fit within the 0051 % FVCOM depths. This means the full profile structure is maintained in 0052 % the resulting FVCOM boundary input despite the differing depths at the 0053 % FVCOM boundary node. 0054 % 2013-06-03 Fix some bugs in the way the open boundary node values were 0055 % stored (the order in which they were stored did not match the order of 0056 % the nodes in Casename_obc.dat). Also fix the order of the vertically 0057 % interpolated values so that FVCOM starts at the surface instead of 0058 % mirroring POLCOMS' approach (where the first value is the seabed). The 0059 % effect of these two fixes (nodes and vertical) should match what FVCOM 0060 % expects. Also add a set of figures (commented out) at the end for 0061 % diagnostic purposes. 0062 % 0063 %========================================================================== 0064 0065 subname = 'get_POLCOMS_tsobc'; 0066 0067 global ftbverbose; 0068 if ftbverbose 0069 fprintf('\n') 0070 fprintf(['begin : ' subname '\n']) 0071 end 0072 0073 if license('test', 'Distrib_Computing_Toolbox') 0074 % We have the Parallel Computing Toolbox, so launch a bunch of workers. 0075 if isempty(gcp('nocreate')) 0076 % Force pool to be local in case we have remote pools available. 0077 parpool('local'); 0078 end 0079 end 0080 0081 varlist = {'lon', 'lat', 'ETWD', 'x1XD', 'time', 'depth', 'pdepthD'}; 0082 0083 % Data format: 0084 % 0085 % pc.ETWD.data and pc.x1XD.data are y, x, sigma, time 0086 % 0087 pc = get_POLCOMS_netCDF(ts, varlist); 0088 0089 [~, ~, nz, nt] = size(pc.ETWD.data); 0090 0091 % Make rectangular arrays for the nearest point lookup. 0092 [lon, lat] = meshgrid(pc.lon.data, pc.lat.data); 0093 0094 %oNodes = Mobj.obc_nodes(Mobj.obc_nodes ~= 0); 0095 % Change the way the nodes are listed to match the order in the 0096 % Casename_obc.dat file. 0097 tmpObcNodes = Mobj.obc_nodes'; 0098 oNodes = tmpObcNodes(tmpObcNodes ~= 0)'; 0099 0100 fvlon = Mobj.lon(oNodes); 0101 fvlat = Mobj.lat(oNodes); 0102 0103 % Number of boundary nodes 0104 nf = sum(Mobj.nObcNodes); 0105 % Number of sigma layers. 0106 fz = size(Mobj.siglayz, 2); 0107 0108 fvtemp = nan(nf, fz, nt); % FVCOM interpolated temperatures 0109 fvsal = nan(nf, fz, nt); % FVCOM interpolated salinities 0110 0111 if ftbverbose 0112 tic 0113 end 0114 for t = 1:nt 0115 if ftbverbose 0116 fprintf('%s : %i of %i timesteps... ', subname, t, nt) 0117 end 0118 % Get the current 3D array of PML POLCOMS-ERSEM results. 0119 pctemp3 = pc.ETWD.data(:, :, :, t); 0120 pcsalt3 = pc.x1XD.data(:, :, :, t); 0121 0122 % Preallocate the intermediate results arrays. 0123 itempz = nan(nf, nz); 0124 isalz = nan(nf, nz); 0125 idepthz = nan(nf, nz); 0126 0127 for j = 1:nz 0128 % Now extract the relevant layer from the 3D subsets. Transpose the 0129 % data to be (x, y) rather than (y, x). 0130 pctemp2 = pctemp3(:, :, j)'; 0131 pcsalt2 = pcsalt3(:, :, j)'; 0132 pcdepth2 = squeeze(pc.depth.data(:, :, j, t))'; 0133 0134 % Create new arrays which will be flattened when masking (below). 0135 tpctemp2 = pctemp2; 0136 tpcsalt2 = pcsalt2; 0137 tpcdepth2 = pcdepth2; 0138 tlon = lon; 0139 tlat = lat; 0140 0141 % Create and apply a mask to remove values outside the domain. This 0142 % inevitably flattens the arrays, but it shouldn't be a problem 0143 % since we'll be searching for the closest values in such a manner 0144 % as is appropriate for an unstructured grid (i.e. we're assuming 0145 % the PML POLCOMS-ERSEM data is irregularly spaced). 0146 mask = tpcdepth2 > 20000; 0147 tpctemp2(mask) = []; 0148 tpcsalt2(mask) = []; 0149 tpcdepth2(mask) = []; 0150 % Also apply the masks to the position arrays so we can't even find 0151 % positions outside the domain, effectively meaning if a value is 0152 % outside the domain, the nearest value to the boundary node will 0153 % be used. 0154 tlon(mask) = []; 0155 tlat(mask) = []; 0156 0157 % Preallocate the intermediate results arrays. 0158 itempobc = nan(nf, 1); 0159 isalobc = nan(nf, 1); 0160 idepthobc = nan(nf, 1); 0161 0162 % Speed up the tightest loop with a parallelized loop. 0163 parfor i = 1:nf 0164 % Now we can do each position within the 2D layer. 0165 0166 fx = fvlon(i); 0167 fy = fvlat(i); 0168 0169 [~, ii] = sort(sqrt((tlon - fx).^2 + (tlat - fy).^2)); 0170 % Get the n nearest nodes from PML POLCOMS-ERSEM data (more? 0171 % fewer?). 0172 ixy = ii(1:16); 0173 0174 % Get the variables into static variables for the 0175 % parallelisation. 0176 plon = tlon(ixy); 0177 plat = tlat(ixy); 0178 ptemp = tpctemp2(ixy); 0179 psal = tpcsalt2(ixy); 0180 pdepth = tpcdepth2(ixy); 0181 0182 % Use a triangulation to do the horizontal interpolation. 0183 tritemp = TriScatteredInterp(plon', plat', ptemp', 'natural'); 0184 trisal = TriScatteredInterp(plon', plat', psal', 'natural'); 0185 triz = TriScatteredInterp(plon', plat', pdepth', 'natural'); 0186 itempobc(i) = tritemp(fx, fy); 0187 isalobc(i) = trisal(fx, fy); 0188 idepthobc(i) = triz(fx, fy); 0189 0190 % Check all three, though if one is NaN, they all will be. 0191 if isnan(itempobc(i)) || isnan(isalobc(i)) || isnan(idepthobc(i)) 0192 warning('FVCOM boundary node at %f, %f is outside the PML POLCOMS-ERSEM domain. Setting to the closest PML POLCOMS-ERSEM value.', fx, fy) 0193 itempobc(i) = tpctemp2(ii(1)); 0194 isalobc(i) = tpcsalt2(ii(1)); 0195 idepthobc(i) = tpcdepth2(ii(1)); 0196 end 0197 end 0198 0199 % Put the results in the intermediate array. 0200 itempz(:, j) = itempobc; 0201 isalz(:, j) = isalobc; 0202 idepthz(:, j) = idepthobc; 0203 0204 end 0205 0206 % Now we've interpolated in space, we can interpolate the z-values 0207 % to the sigma depths. 0208 0209 % Preallocate the output arrays 0210 fvtempz = nan(nf, fz); 0211 fvsalz = nan(nf, fz); 0212 0213 for pp = 1:nf 0214 % Get the FVCOM depths at this node 0215 tfz = Mobj.siglayz(oNodes(pp), :); 0216 % Now get the interpolated PML POLCOMS-ERSEM depth at this node 0217 tpz = idepthz(pp, :); 0218 0219 % To ensure we get the full vertical expression of the vertical 0220 % profiles, we need to normalise the POLCOMS-ERSEM and FVCOM 0221 % depths to the same range. This is because in instances where 0222 % FVCOM depths are shallower (e.g. in coastal regions), if we 0223 % don't normalise the depths, we end up truncating the vertical 0224 % profile. This approach ensures we always use the full 0225 % vertical profile, but we're potentially squeezing it into a 0226 % smaller depth. 0227 A = max(tpz); 0228 B = min(tpz); 0229 C = max(tfz); 0230 D = min(tfz); 0231 norm_tpz = (((D - C) * (tpz - A)) / (B - A)) + C; 0232 0233 % Get the temperature and salinity values for this node and 0234 % interpolate down the water column (from PML POLCOMS-ERSEM to 0235 % FVCOM). I had originally planned to use csaps for the 0236 % vertical interplation/subsampling at each location. However, 0237 % the demo of csaps in the MATLAB documentation makes the 0238 % interpolation look horrible (shaving off extremes). interp1 0239 % provides a range of interpolation schemes, of which pchip 0240 % seems to do a decent job of the interpolation (at least 0241 % qualitatively). 0242 if ~isnan(tpz) 0243 fvtempz(pp, :) = interp1(norm_tpz, itempz(pp, :), tfz, 'pchip', 'extrap'); 0244 fvsalz(pp, :) = interp1(norm_tpz, isalz(pp, :), tfz, 'pchip', 'extrap'); 0245 else 0246 warning('Should never see this... ') % because we test for NaNs when fetching the values. 0247 warning('FVCOM boundary node at %f, %f is outside the PML POLCOMS-ERSEM domain. Skipping.', fvlon(pp), fvlat(pp)) 0248 continue 0249 end 0250 end 0251 0252 % The horizontally- and vertically-interpolated values in the final 0253 % FVCOM results array. 0254 fvtemp(:, :, t) = fvtempz; 0255 fvsal(:, :, t) = fvsalz; 0256 0257 if ftbverbose 0258 fprintf('done.\n') 0259 end 0260 end 0261 if ftbverbose 0262 toc 0263 end 0264 0265 Mobj.temperature = fvtemp; 0266 Mobj.salt = fvsal; 0267 0268 % Convert the current times to Modified Julian Day (this is a bit ugly). 0269 pc.time.all = strtrim(regexp(pc.time.units, 'since', 'split')); 0270 pc.time.datetime = strtrim(regexp(pc.time.all{end}, ' ', 'split')); 0271 pc.time.ymd = str2double(strtrim(regexp(pc.time.datetime{1}, '-', 'split'))); 0272 pc.time.hms = str2double(strtrim(regexp(pc.time.datetime{2}, ':', 'split'))); 0273 0274 Mobj.ts_times = greg2mjulian(... 0275 pc.time.ymd(1), ... 0276 pc.time.ymd(2), ... 0277 pc.time.ymd(3), ... 0278 pc.time.hms(1), ... 0279 pc.time.hms(2), ... 0280 pc.time.hms(3)) + (pc.time.data / 3600 / 24); 0281 0282 if ftbverbose 0283 fprintf(['end : ' subname '\n']) 0284 end 0285 0286 0287 %% 0288 % Plot a vertical profile for a boundary node (for my Irish Sea case, this 0289 % is one of the ones along the Celtic Sea boundary). Also plot the 0290 % distribution of interpolated values over the POLCOMS data. Add the 0291 % location of the vertical profile (both FVCOM and POLCOMS) to the plot. 0292 % nn = 55; % open boundary index 0293 % tt = 1; % time index 0294 % 0295 % % Get the corresponding indices for the POLCOMS data 0296 % [~, xidx] = min(abs(lon(1, :) - fvlon(nn))); 0297 % [~, yidx] = min(abs(lat(:, 1) - fvlat(nn))); 0298 % 0299 % % close all 0300 % 0301 % figure 0302 % clf 0303 % subplot(2,2,1) 0304 % plot(Mobj.temperature(nn, :, 1), Mobj.siglayz(oNodes(nn), :), 'x-') 0305 % xlabel('Temperature (^{\circ}C)') 0306 % ylabel('Depth (m)') 0307 % title('FVCOM') 0308 % 0309 % subplot(2,2,2) 0310 % % Although POLCOMS stores its temperature values from seabed to surface, 0311 % % the depths are stored surface to seabed. Nice. 0312 % plot(squeeze(pc.ETWD.data(xidx, yidx, :, 1)), squeeze(pc.depth.data(xidx, yidx, :, 1)), 'rx-') 0313 % xlabel('Temperature (^{\circ}C)') 0314 % ylabel('Depth (m)') 0315 % title('POLCOMS') 0316 % 0317 % subplot(2,2,3) 0318 % plot(Mobj.temperature(nn, :, tt), 1:fz, 'x-') 0319 % xlabel('Temperature (^{\circ}C)') 0320 % ylabel('Array index') 0321 % title('FVCOM') 0322 % 0323 % subplot(2,2,4) 0324 % plot(squeeze(pc.ETWD.data(xidx, yidx, :, tt)), 1:nz, 'rx-') 0325 % xlabel('Temperature (^{\circ}C)') 0326 % ylabel('Array index') 0327 % title('POLCOMS') 0328 % 0329 % % Figure to check everything's as we'd expect. Plot first time step with 0330 % % the POLCOMS surface temperature as a background with the interpolated 0331 % % boundary node surface values on top. 0332 % 0333 % figure 0334 % clf 0335 % % Plot POLCOMS surface data (last sigma layer) 0336 % dx = mean(diff(pc.lon.data)); 0337 % dy = mean(diff(pc.lat.data)); 0338 % pcolor(pc.lon.data - (dx / 2), pc.lat.data - (dy / 2), ... 0339 % squeeze(pc.ETWD.data(:, :, 1, tt))') 0340 % shading flat 0341 % axis('equal', 'tight') 0342 % daspect([1.5, 1, 1]) 0343 % hold on 0344 % % Add the interpolated surface data (first sigma layer) 0345 % scatter(Mobj.lon(oNodes), Mobj.lat(oNodes), 40, Mobj.temperature(:, 1, tt), 'filled', 'MarkerEdgeColor', 'k') 0346 % axis([min(Mobj.lon(oNodes)), max(Mobj.lon(oNodes)), min(Mobj.lat(oNodes)), max(Mobj.lat(oNodes))]) 0347 % caxis([6, 20]) 0348 % plot(lon(yidx, xidx), lat(yidx, xidx), 'rs') % polcoms is all backwards 0349 % plot(Mobj.lon(oNodes(nn)), Mobj.lat(oNodes(nn)), 'wo') 0350 % colorbar