Distributes precipitation on the preceeding timesteps in a physically plausible way.
This assumes that the precipitation has been measured on intervals greater than the sampling interval of the data file (for example, 24 hours accumulations written once per day in an hourly file, the other timesteps receiving nodata). And the accumulation has to be written on the last timestep of the accumulation period.
It supports the following arguments:
- SOFT: if set to TRUE, missing accumulated values would be replaced by "0";
- MEAS_PERIOD: measured accumulation period in seconds (mandatory).
The precipitation is distributed on the preceeding timesteps by using criterias on relative humidity and the difference between the air temperature and the surface temperature.
PSUM::filter1 = PSUM_DISTRIBUTE
PSUM::arg1::meas_period = 86400
|
| ProcPSUMDistribute (const std::vector< std::pair< std::string, std::string > > &vecArgs, const std::string &name, const Config &cfg) |
|
virtual void | process (const unsigned int ¶m, const std::vector< MeteoData > &ivec, std::vector< MeteoData > &ovec) |
| Distributes a precipitation sum over the most probable preceeding time steps. This works by looking at various criterias (like relative humidity, TA-TSS, etc) and picking up the time steps that show the highest scores. About this index usage in the code below: When looking into the ovec vector, they are: More...
|
|
virtual | ~ProcessingBlock () |
|
virtual void | process (const unsigned int ¶m, const std::vector< MeteoData > &ivec, std::vector< MeteoData > &ovec)=0 |
|
virtual void | process (Date &dateStart, Date &dateEnd) |
|
std::string | getName () const |
|
const ProcessingProperties & | getProperties () const |
|
const std::string | toString () const |
|
bool | skipStation (const std::string &station_id) const |
| Should the provided station be skipped in the processing? More...
|
|
bool | noStationsRestrictions () const |
|
const std::vector< DateRange > | getTimeRestrictions () const |
|
bool | skipHeight (const double &height) const |
| Should the provided height be skipped in the processing? More...
|
|
|
static void | SmartDistributePSUM (const double &precip, const size_t &start_idx, const size_t &end_idx, const size_t ¶mindex, std::vector< MeteoData > &vecM) |
|
static void | CstDistributePSUM (const double &precip, const size_t &start_idx, const size_t &end_idx, const size_t ¶mindex, std::vector< MeteoData > &vecM) |
|
static void | readCorrections (const std::string &filter, const std::string &filename, std::vector< double > &X, std::vector< double > &Y) |
| Read a data file structured as X Y value on each lines. More...
|
|
static void | readCorrections (const std::string &filter, const std::string &filename, std::vector< double > &X, std::vector< double > &Y1, std::vector< double > &Y2) |
| Read a data file structured as X Y1 Y2 value on each lines. More...
|
|
static std::vector< double > | readCorrections (const std::string &filter, const std::string &filename, const size_t &col_idx, const char &c_type, const double &init) |
| Read a correction file applicable to repeating time period. More...
|
|
static std::vector< offset_spec > | readCorrections (const std::string &filter, const std::string &filename, const double &TZ, const size_t &col_idx=2) |
| Read a correction file, ie a file structured as timestamps followed by values on each lines. More...
|
|
static std::map< std::string, std::vector< DateRange > > | readDates (const std::string &filter, const std::string &filename, const double &TZ) |
| Read a list of date ranges by stationIDs from a file. More...
|
|
void mio::ProcPSUMDistribute::process |
( |
const unsigned int & |
param, |
|
|
const std::vector< MeteoData > & |
ivec, |
|
|
std::vector< MeteoData > & |
ovec |
|
) |
| |
|
virtual |
Distributes a precipitation sum over the most probable preceeding time steps. This works by looking at various criterias (like relative humidity, TA-TSS, etc) and picking up the time steps that show the highest scores. About this index usage in the code below: When looking into the ovec vector, they are:
___▒_____________▒_____________▒____
↑ ↑
ii endIdx
The goal is that the only "hard" references we can have are the accumulated value (ie the only values != nodata) and therefore their index are always kept (as ii and endIdx). This leads to the many "ii+1" in the code below since the period over which the data has to be distributed starts at ii+1 and goes until endIdx (which will become the new ii for the next iteration).
Implements mio::ProcessingBlock.