2

更新:更改了这个问题以更好地反映我目前的理解。

我有一个 NetCDF 4.5 版 Grib2Record 对象。给定一个 (x,y) 网格点和一个变量名称,我想按预测时间从对象中提取该变量的所有预测数据(如果记录包含该变量的预测)。由于写入磁盘索引文件的默认行为,我不想使用更高级别的 NetCDFFile 接口。

我曾尝试查看较低级别的代码(Grib2Rectilyser、Grib2Customizer 等),但代码太密集了,我正在寻找从哪里开始的帮助。

我将不胜感激有关如何获取 Grib2Record 和 1. 检查特定预测变量是否包含在其中的任何指示,以及 2. 如果是,则通过给定 xy 网格点的预测有效时间提取预测数据和z 级。

4

1 回答 1

5

我使用 grib2 文件进行风预测,这就是我获取记录的方式以及如何处理它以获取风(VU 组件)

Grib2Input input = new Grib2Input(getRandomAccessFile());
if (!input.scan(false, false)) {
    logger.error("Failed to successfully scan grib file");
    return;
}
Grib2Data data = new Grib2Data(getRandomAccessFile());

List<Grib2Record> records = input.getRecords();

for (Grib2Record record : records) {    
    Grib2IndicatorSection is = record.getIs();
    Grib2IdentificationSection id = record.getId();
    Grib2Pds pdsv = record.getPDS().getPdsVars();
    Grib2GDSVariables gdsv = record.getGDS().getGdsVars();

    long time = id.getRefTime() + (record.getPDS().getForecastTime() * 3600000);

    logger.debug("Record description at " + pdsv.getReferenceDate() + " forecast "
    + new Date(time)    + ": " + ParameterTable.getParameterName(is.getDiscipline(), pdsv.getParameterCategory(), pdsv.getParameterNumber()));

    float[] values = data.getData(record.getGdsOffset(), record.getPdsOffset(), 0);

     if ((is.getDiscipline() == 0) && (pdsv.getParameterCategory() == 2) && (pdsv.getParameterNumber() == 2)) {
        // U-component_of_wind
        int c = 0;
        for (double lat = gdsv.getLa1(); lat >= gdsv.getLa2(); lat = lat - gdsv.getDy()) {
            for (double lon = gdsv.getLo1(); lon <= gdsv.getLo2(); lon = lon + gdsv.getDx()) {
                logger.debug(lat + "\t" + lon + "\t" +
                values[c]);
                c++;
            }
        }
    } else if ((is.getDiscipline() == 0) && (pdsv.getParameterCategory() == 2) && (pdsv.getParameterNumber() == 3)) {
        // V-component_of_wind
        int c = 0;
        for (double lat = gdsv.getLa1(); lat >= gdsv.getLa2(); lat = lat - gdsv.getDy()) {
            for (double lon = gdsv.getLo1(); lon <= gdsv.getLo2(); lon = lon + gdsv.getDx()) {
                logger.debug(lat + "\t" + lon + "\t" +
                values[c]);
                c++;
            }
        }
    }
}

private RandomAccessFile getRandomAccessFile() {
    RandomAccessFile raf = null;
    try {
        raf = new RandomAccessFile(path, "r");
        raf.order(RandomAccessFile.BIG_ENDIAN);
    } catch (IOException e) {
        logger.error("Error opening file " + path, e);
    }
    return raf;
}
于 2015-01-15T16:57:28.213 回答