Data Description

Parameters

SMAP Level-4 soil moisture data include the following parameters:

• Surface soil moisture (0-5 cm vertical average)
• Root zone soil moisture (0-100 cm vertical average)
• Additional research products (not validated), including surface meteorological forcing variables, soil temperature, evapotranspiration, net radiation, and error estimates for select output fields that are produced internally by the SMAP Level-4 soil moisture algorithm

Soil moisture is output in volumetric units, in wetness (or relative saturation) units, and in percentile units (except surface soil moisture). 

Refer to the Appendix of this document for details on all parameters. Parameters are further described in the Algorithm Theoretical Basis Document (ATBD) for this product under Section 3: Physics of the Problem (Reichle et al. 2014).

File Information

Format

Data are in HDF5 format. For software and more information, including an HDF5 tutorial, visit the HDF Group's HDF5 website.

File Contents

SMAP Level-4 soil moisture data consists of three main products:

• Geophysical Data
• Analysis Update Data
• Land Model Constants

For each 3-hour interval, there are two files: one geophysical (gph) file and one analysis update (aup) file. Land model constants (lmc) are provided in a single file per Science Version. Science Version IDs (such as Vv3030) are included in all file names, and are defined in the File Naming Convention section of this user guide.

• Analysis Data: Soil moisture and temperature analysis estimates, including error estimates
• Forecast Data: Land model predictions of brightness temperature, soil moisture, and soil temperature
• Observations Data: Assimilated SMAP brightness temperature observations and data assimilation diagnostics

Figure 2 shows a subset of the aup file contents.

Data Fields

All global data arrays are two dimensional with 1624 rows and 3856 columns (6,262,144 pixels per layer).

Metadata Fields

Each product also contains metadata that describe the full content of each file. For a description of all metadata fields for this product, refer to the Product Specification Document.

Naming Convention

Files are named according to the following convention, which is described in Table 1: 

SMAP_L4_SM_pid_yyyymmddThhmmss_VLMmmm_NNN.[ext]

For example:

SMAP_L4_SM_gph_20151015T133000_Vv3030_001.h5

Where:

File Size

Table 2 provides file sizes and daily volume estimates for each product.

Spatial Information

Coverage

Coverage spans from 180°W to 180°E, and from approximately 85.044°N to 85.044°S. The gap in coverage at both the North and South Pole, called a pole hole, has a radius of approximately 400 km. Coverage is for the global land surface excluding inland water and permanently frozen areas.

Resolution

The native spatial resolution of the radiometer footprint is approximately 40 km. Data are then assimilated into a land surface model that is gridded using the 9 km global EASE-Grid 2.0 projection.

Projection and Grid Information

These data are provided on the global cylindrical EASE-Grid 2.0 (Brodzik et al. 2012). Each grid cell has a nominal area of approximately 81 km² regardless of longitude and latitude.

EASE-Grid 2.0 has a flexible formulation. By adjusting a single scaling parameter, a family of multi-resolution grids that nest within one another can be generated. The nesting can be adjusted so that smaller grid cells can be tessellated to form larger grid cells. Figure 4 shows a schematic of the nesting to a resolution of 3 km (4872 rows x 11568 columns on global coverage), 9 km (1624 rows x 3856 columns on global coverage) and 36 km (406 rows x 964 columns on global coverage).

This feature of perfect nesting provides SMAP data products with a convenient common projection for both high-resolution radar observations and low-resolution radiometer observations, as well as for their derived geophysical products.

For more on EASE-Grid 2.0, refer to the EASE Grids website.

Temporal Information

Coverage

Coverage is continuous and spans from 31 March 2015 to present.

SMAP Satellite and Processing Events

Due to instrument maneuvers, data downlink anomalies, data quality screening, and other factors, small gaps in the SMAP time series will occur. Details of these events are maintained on two master lists:

• SMAP On-Orbit Events List for Instrument Data Users
• Master List of Bad and Missing Data

However, gaps in the SMAP time series do not affect this product. While some temporal coverage gaps exist in the input SPL1CTB data, the SPL4SM product is processed continuously and does not have temporal coverage gaps. When SPL1CTB gaps occur, SPL4SM data are processed using information from SMAP observations assimilated prior to each gap in the input SPL1CTB data, as well as information from the land surface model.

For the period between 19 June and 23 July 2019, an extended gap occurred in the L1 - L3 SMAP products. During this period, the L4 Soil Mositure data sets were not informed by SMAP data. For more information on this SMAP outage, users should refer to the SMAP Post-Recovery Notice. 

Latencies

FAQ: What are the latencies for SMAP radiometer data sets?

Resolution

Three basic time steps are involved in the generation of the Level-4 soil moisture products, including:

1. The land model computational time step (7.5 minutes)
2. The Ensemble Kalman Filter (EnKF) analysis update time step (3 hours)
3. The reporting/output time step for the instantaneous and time-average geophysical fields that are stored in the data products (3 hours)

SMAP observations are assimilated in an EnKF analysis update step at the nearest 3-hourly analysis time such as 0z, 3z, ..., and 21z (where z indicates Zulu). A broad variety of geophysical parameters are provided as 3-hourly averages between these update times. Moreover, instantaneous forecast and analysis soil moisture and temperature estimates are provided along with the assimilated observations. These snapshots are nominally for 0z, 3z,…, or 21z.

Data Acquisition and Processing

This section has been adapted from Reichle et al. 2014, the ATBD for this product. Additional documentation of the algorithm is provided by Reichle et al. 2017a, Reichle et al. 2017b, and Reichle et al. 2019.

Background

The primary SMAP measurements, land surface microwave emission at 1.41 GHz and radar backscatter at 1.26 GHz and 1.29 GHz, are directly related to surface soil moisture (in the top 5 cm of the soil column). Several of the key applications targeted by SMAP, however, require knowledge of root zone soil moisture (defined here as soil moisture in the top 1 m of the soil column), which is not directly linked to SMAP observations. The foremost objective of the SMAP Level-4 Surface and Root Zone Soil Moisture (SPL4SM) product is to fill this gap and provide estimates of root zone soil moisture that are informed by and consistent with SMAP observations. Such estimates are obtained by merging SMAP observations with estimates from a land surface model in a soil moisture data assimilation system.

The land surface model component of the assimilation system is driven with observation-based surface meteorological forcing data, including precipitation, which is the most important driver for soil moisture. The model also encapsulates knowledge of key land surface processes, including the vertical transfer of soil moisture between the surface and root zone reservoirs. Finally, the assimilation system uses the land model to interpolate and extrapolate SMAP observations in time and in space. The SPL4SM product thus provides a comprehensive and consistent picture of land surface hydrological conditions based on SMAP observations and complementary information from a variety of sources. The assimilation algorithm considers the respective uncertainties of each component and, if properly calibrated, yields a product that is superior to both satellite and land model data. Error estimates for the SPL4SM product are generated as a by-product of the data assimilation system.

The ATBD for this product provides a detailed description of the SPL4SM product, its algorithm, and how the product is validated.

Acquisition

SMAP Level-4 soil moisture products are derived from the following data sets: 

• SMAP L1C Radiometer Half-Orbit 36 km EASE-Grid Brightness Temperatures, Version 4 (SPL1CTB)
• GEOS-5 Forward Processing (FP) Model Data from the NASA Global Modeling and Assimilation Office (GMAO): Daily surface meteorology from observation-constrained global model analysis; includes precipitation corrections using the NOAA Climate Prediction Center "Unified" global, 0.5 degree, daily gauge-based data product. (Reichle et al. 2017a, Reichle et al. 2017b, Reichle et al. 2019)

In addition, ancillary data sources used as input to calculating the SMAP Level-4 soil moisture products are obtained from the GMAO; these sources are listed in the ATBD, Section 4.1.3: Ancillary Data Requirements. Precipitation observations that are used to correct the GMAO precipitation estimates are obtained from the NOAA Climate Prediction Center (Reichle et al. 2017a, Reichle et al. 2017b, Reichle et al. 2019).

Utilizing the baseline data assimilation algorithm discussed below, input data sources are used with the SMAP Level-4 soil moisture model to provide enhanced estimates of surface soil moisture, root zone soil moisture, and related geophysical variables.

Baseline Algorithm

The SPL4SM science algorithm consists of two key processing elements:

1. GEOS-5 Catchment Land Surface and Microwave Radiative Transfer Model
2. GEOS-5 Ensemble-Based Land Data Assimilation Algorithm

The GEOS-5 Catchment Land Surface and Microwave Radiative Transfer Model is a numerical description of the water and energy transport processes at the land-atmosphere interface, augmented with a model that describes the land surface microwave radiative transfer (refer to section 4.1.1 of the ATBD: Reichle et al. 2014). The GEOS-5 Ensemble-Based Land Data Assimilation System is the tool used to merge SMAP observations with estimates from the land model as it is driven with observation-based surface meteorological forcing data.

The SMAP Level-4 soil moisture baseline algorithm, described in detail in the ATBD, includes a soil moisture analysis based on the ensemble Kalman filter and a rule-based freeze/thaw analysis. However, data users should note that for Validated Version 4 data, the algorithm ingests only the SPL1CTB radiometer brightness temperatures, contrary to the planned use of downscaled brightness temperatures from the SPL2SMAP product and of landscape freeze-thaw state retrievals from the SPL2SMA product. The latter two products—SPL2SMAP and SPL2SMA—are based on radar observations and are only available for the period from 13 April 2015 through 07 July 2015 due to an anomaly that caused the premature failure of the SMAP L-band radar. Neither of these two radar-based products is assimilated in the SMAP Level-4 soil moisture algorithm.

More information about error sources is provided in the ATBD under Section 4.1.2: Mathematical Description of the Algorithm. For more information on data product accuracy, refer to Reichle et al. 2017a, Reichle et al. 2017b, Reichle et al. 2019, and the Validated Assessment Report from Reichle et al. 2018.

Processing

SMAP Level-4 soil moisture data are generated by the GMAO located at the NASA Goddard Space Flight Center (GSFC), using the High-End Computing Facilities at the NASA Center for Climate Simulation (NCCS), also located at GSFC in Greenbelt, Maryland.

SMAP SPL1CTB data are required for the baseline algorithm. Aside from SMAP observations, the data assimilation system requires initialization, parameter, and forcing inputs for the Catchment land surface model, as well as input error parameters for the ensemble-based data assimilation system. These ancillary data requirements are described in detail in the ATBD, Section 4.1.3: Ancillary Data Requirements. The precipitation observations used to correct the GMAO precipitation estimates are obtained from the NOAA Climate Prediction Center (Reichle et al. 2017a, Reichle et al. 2017b, Reichle et al. 2019). Note that for this version, the model background precipitation forcing is rescaled to match the climatology of the Global Precipitation Climatology Project (v2.2), which results in substantial changes in the precipitation and soil moisture climatology in Africa and the high latitudes, where the gauge-based Climate Prediction Center Unified precipitation is not used.

For more information on each portion of the algorithm processing flow, refer to the ATBD.

Land Surface Modeling System and SMAP Nature Run 

Note that for Version 4 SPL4SM products an improved version of the land surface modeling system is used. The corresponding model-only Nature Run (NRv7.2) simulation is used to derive brightness temperature scaling parameters, model soil moisture initial conditions, and the soil moisture climatology. For this release, the land surface modeling system was revised in the following ways:

• Improved input parameter data sets for land cover, topography, and vegetation height are based on more recent data sets. Land cover inputs were updated to the GlobCover2009 product, resulting in a slightly different land mask between Version 3 and Version 4. Topographic statistics now rely on observations from the Shuttle Radar Topography Mission. Finally, vegetation height inputs are derived from space-borne lidar measurements.  
• The model background precipitation forcing is rescaled to match the climatology of the Global Precipitation Climatology Project (v2.2), which results in substantial changes in the precipitation and soil moisture climatology in Africa and the high latitudes, where the gauge-based Climate Prediction Center Unified precipitation is not used.
• SMAP Level-2 soil moisture retrievals and in situ soil moisture measurements from the Soil Climate Analysis Network and U.S. Climate Reference Network were used to calibrate a particular Catchment model parameter that governs the recharge of soil moisture from the model’s root-zone excess reservoir into the surface excess reservoir. Specifically, the replenishment of soil moisture near the surface from below under non-equilibrium conditions was substantially reduced, which brings the model’s surface soil moisture more in line with the SMAP Level-2 and in situ soil moisture.  
• Additional model changes include revisions to the parameters and parameterizations of the surface energy balance and the snow depletion curve.

Quality, Errors, and Limitations

Quality Assessments

For in-depth details regarding the quality of these data, refer to the Validated Assessment Report.

Quality Overview

SMAP products provide multiple means to assess quality. Uncertainty measures and file-level metadata that provide quality information are provided within each product. For details, refer to the Appendix of this document and the Product Specification Documents.

Each HDF5 file contains file-level metadata. A separate metadata file with an .xml file extension is available from the NSIDC DAAC with every HDF5 file; it contains essentially the same information as the file-level metadata. In addition, a Quality Assessment (QA) file with a .qa file extension is provided for every HDF5 file. QA files contain spatial statistics across the SMAP Level-4 soil moisture products, such as the global minimum, mean, and maximum of each data field.

Level-4 surface and root zone soil moisture estimates are validated to a Root Mean Square Error (RMSE) requirement of 0.04 m³ m^-3 after removal of the long-term mean bias. This accuracy requirement is identical to Level-2 soil moisture product validation and excludes regions with snow and ice cover, frozen ground, mountainous topography, open water, urban areas, and vegetation with water content greater than 5 kg m^-2. Research outputs (not validated) include the surface meteorological forcing fields, land surface fluxes, soil temperature and snow states, runoff, and error estimates that are derived from the ensemble.

Quality Control

Quality control is also an integral part of the soil moisture assimilation system. Two kinds of quality control (QC) measures are applied. The first set of QC steps is based on the flags that are provided with the SMAP observations. Only SMAP brightness temperature data that have favorable flags for soil moisture estimation are assimilated, such as acceptably low vegetation density, no rain, no snow cover, no frozen ground, no RFI, sufficient distance from open water, etc.

The second set of QC steps are additional rules that exclude SMAP observations from assimilation in the EnKF soil moisture update whenever the land surface model indicates that (1) heavy rain is falling, (2) the soil is frozen, or (3) the ground is fully or partly covered with snow. The assimilation system will typically provide some weight to the model background and thus buffers the impact of anomalous observations that are not caught in the flagging process.

Note: Brightness temperature observations from Version 4 SPL1CTB granules that have known deficiencies were excluded from assimilation in the Version 4 SPL4SM algorithm.

For more quality control information, refer to the Appendix of this doocument.

Error Sources

The data assimilation system weighs the relative errors of the assimilated lower-level product (such as radiance or retrieval) and the land model forecast. Estimates of the error of the assimilation product are dynamically determined as a by-product of this calculation. How useful these error estimates are depends on the accuracy of the input error parameters and needs to continue to be determined through validation; refer to the ATBD, Section 4.2.4. The target accuracy of the assimilated brightness temperatures is discussed in the SPL1CTB product documentation. Error estimates of the land surface model and required input error parameters are discussed in the ATBD for this product.

Each instantaneous land model field is accompanied with a corresponding instantaneous error field which is provided for select variables. The relevant outputs are listed in the Data Fields document for the SPL4SMAU product. Specifically, the error estimates are derived from the ensemble standard deviation of the analyzed fields. For soil moisture, the ensemble standard deviation is computed from the analysis ensemble in volumetric units (m³ m^-3). For temperatures, the ensemble standard deviation is provided in units of kelvin. These error estimates will vary in space and time.

More information about error sources is provided in the ATBD under Section 4.1.2: Mathematical Description of the Algorithm. For more information on data product accuracy, refer to Reichle et al. 2017a, Reichle et al. 2017b, Reichle et al. 2019, and the Validated Assessment Report from Reichle et al. 2018.

Instrumentation

For a detailed description of the SMAP instrument, visit the SMAP Instrument page at the JPL SMAP website.

Software and Tools

For tools that work with SMAP data, refer to the Tools web page.

Version History

Related Data Sets

SMAP Data at NSIDC | Overview

SMAP Radar Data at the ASF DAAC

Related Websites

SMAP at NASA JPL

Contacts and Acknowledgments

Investigators

Rolf H. Reichle, Randal Koster, Qing Liu
NASA Goddard Space Flight Center
Greenbelt, MD

Gabrielle De Lannoy
KU Leuven
Department of Earth and Environmental Sciences
Heverlee, Belgium

Wade Crow
Hydrology and Remote Sensing Lab
US Department of Agriculture/Agricultural Research Service (USDA ARS)
Beltsville, MD

John Kimball
Numerical Terradynamic Simulation Group
University of Montana
Missoula, MT

References

References

Brodzik, M. J., B. Billingsley, T. Haran, B. Raup, and M. H. Savoie. 2012. EASE-Grid 2.0: Incremental but Significant Improvements for Earth-Gridded Data Sets. ISPRS Int. J. Geo-Inf. 1(1):32-45. https://dx.doi.org/10.3390/ijgi1010032.

Brodzik, M. J., B. Billingsley, T. Haran, B. Raup, and M. H. Savoie. 2014. Correction: Brodzik, M. J. et al. EASE-Grid 2.0: Incremental but Significant Improvements for Earth-Gridded Data Sets. ISPRS Int. J. Geo-Inf 2012. 1(1):32-45 ISPRS Int. J. Geo-Inf. 3(3):1154-1156. https://dx.doi.org/10.3390/ijgi3031154.

Crow, W. T., F. Chen, R. H. Reichle, and Q. Liu. 2017. L Band Microwave Remote Sensing and Land Data Assimilation Improve the Representation of Prestorm Soil Moisture Conditions for Hydrologic Forecasting. Geophysical Research Letters. 44:5495-5503. https://dx.doi.org/10.1002/2017GL073642.

Crow, W. T., F. Chen, R. H. Reichle, Y. Xia, and Q. Liu. 2018. Exploiting soil moisture, precipitation and streamflow observations to evaluate soil moisture/runoff coupling in land surface models. Geophysical Research Letters. 45, in press, https://doi.org/10.1029/2018GL077193.

De Lannoy, G. J. M., and R. H. Reichle. 2016. Assimilation of SMOS Brightness Temperatures or Soil Moisture Retrievals into a Land Surface Model. Hydrology and Earth System Sciences. 20:4895-4911. Hydrol. Earth Syst. Sci., 20:4895-4911. https://dx.doi.org/10.5194/hess-20-4895-2016.

De Lannoy, G. J. M., and R. H. Reichle. 2016. Global Assimilation of Multiangle and Multipolarization SMOS Brightness Temperature Observations into the GEOS-5 Catchment Land Surface Model for Soil Moisture Estimation. Journal of Hydrometeorology, 17:669-691. https://dx.doi.org/10.1175/JHM-D-15-0037.1.

Entekhabi, D., R. H. Reichle, R. D. Koster, and W. T. Crow. 2010. Performance Metrics for Soil Moisture Retrievals and Application Requirements. Journal of Hydrometeorology. 11:832–840. https://dx.doi.org/10.1175/2010JHM1223.1.

Koster, R. D., Q. Liu, S. P. P. Mahanama, and R. H. Reichle. 2018. Improved Hydrological Simulation Using SMAP Data: Relative Impacts of Model Calibration and Data Assimilation. Journal of Hydrometeorology. In press. https://dx.doi.org/10.1175/JHM-D-17-0228.1.

Reichle, R. H., and Q. Liu. 2014. Observation-Corrected Precipitation Estimates in GEOS-5. SMAP Project, Global Modeling and Assimilation Office, Goddard Space Flight Center, Greenbelt, MD, USA. NASA/TM–2014-104606, Vol. 35. (https://gmao.gsfc.nasa.gov/pubs/docs/Reichle734.pdf, 495 KB)

Reichle, R. et al. 2014. SMAP Algorithm Theoretical Basis Document: L4 Surface and Root-Zone Soil Moisture Product. SMAP Project, JPL D-66483, Jet Propulsion Laboratory, Pasadena, CA, USA. (PDF, 1.4 MB; see Technical References)

Reichle, R., G. J. M. De Lannoy, Q. Liu, J. V. Ardizzone, F. Chen, A. Colliander, A. Conaty, W. Crow, T. Jackson, J. Kimball, R. D. Koster, and E. Brent Smith. 2016. Soil Moisture Active Passive Mission L4_SM Data Product Assessment (Version 2 Validated Release). GMAO Office Note No. 12 (Version 1.0), 55 pp, NASA Goddard Space Flight Center, Greenbelt, MD, USA. (PDF, 3.2 MB; see Technical References)

Reichle, R. H., et al.  2017a. Assessment of the SMAP Level-4 Surface and Root-Zone Soil Moisture Product Using In Situ Measurements. Journal of Hydrometeorology 18:2621-2645. http://dx.doi.org/doi:10.1175/JHM-D-17-0063.1. 

Reichle, R. H., G. J. De Lannoy, Q. Liu, R. D. Koster, J. S. Kimball, W. T. Crow, J. V. Ardizzone, P. Chakraborty, D. W. Collins, A. L. Conaty, M. Girotto, L. A. Jones, J. Kolassa, H. Lievens, R. A. Lucchesi, and E. B. Smith. 2017b. Global Assessment of the SMAP Level-4 Surface and Root-Zone Soil Moisture Product Using Assimilation Diagnostics. Journal of Hydrometeorology, accepted. https://doi.org/10.1175/JHM-D-17-0130.1.

Reichle, R. H., Q. Liu, R. D. Koster, J. Ardizzone, A. Colliander, W. Crow, G. J. M. De Lannoy, and J. Kimball. 2018. Soil Moisture Active Passive (SMAP) Project Assessment Report for Version 4 of the L4_SM Data Product. National Aeronautics and Space Administration: Technical Report Series on Global Modeling and Data Assimilation, Volume 52. (PDF, 2.6 MB; see Technical References)

Reichle, R. H., Q. Liu, R. D. Koster, W. T. Crow, G. J. M. De Lannoy, J. S. Kimball, J. V. Ardizzone, ... and J. P. Walker. 2019. Version 4 of the SMAP Level‐4 Soil Moisture Algorithm and Data Product. J. of Advances in Modeling Earth Systems, in press. https://doi.org/10.1029/2019MS001729.

Appendix - Data Fields

This appendix provides a description of all data fields within the SMAP L4 Global 3-hourly 9 km Surface and Rootzone Soil Moisture Geophysical Data, SMAP L4 Global 3-hourly 9 km EASE-Grid Surface and Root Zone Soil Moisture Analysis Update, and SMAP L4 Global 9 km EASE-Grid Surface and Root Zone Soil Moisture Land Model Constants products. The data are grouped into the following main HDF5 groups: 

• Geophysical Data (gph)*
• Analysis Data (aup)
• Forecast_Data (aup)
• Observations_Data (aup)
• Land-Model-Constants Data (lmc)
• Metadata

For a description of metadata fields for this product, refer to the Product Specification Document.

* As reflected in the file names, gph, aup, and lmc indicate three different file collections: geophysical, analysis, and land-model-constants data, respectively. Note that analysis, forecast, and observations data are contained in the aup collection.

Geophysical_Data

Table A1 describes the data fields in the Geophysical_Data group stored in the gph file collection. This group contains fields that specify time-average geophysical data (including soil moisture, soil temperature, and land surface fluxes). Time and space coordinate information is stored in the HDF5 root data group.

Analysis_Data

Table A2 describes the data fields in the Analysis_Data group stored in the aup file collection. This group contains soil moisture and temperature estimates after the ensemble Kalman filter analysis update, along with their corresponding uncertainty estimates. Soil moisture and temperature values are snapshots/instantaneous data. Time and space coordinate information is stored in the HDF5 root data group.

Forecast_Data

Table A3 describes the data fields in the Forecast_Data group stored in the aup file collection. This group is the land model equivalent of the Observations_Data group; it provides the land surface model’s predictions of the assimilated observations. These forecasts, or observation predictions, are based on propagating the land surface model forward in time from the previous analysis time step. The Forecast_Data group does not contain a medium-range (5-day) forecast of land surface conditions. Soil moisture and temperature values are snapshots/instantaneous data. Time and space coordinate information is stored in the HDF5 root data group.

Observations_Data

Table 4 describes the data fields in the Observations_Data group stored in the aup file collection. This group provides information about the assimilated SMAP observations. Time and space coordinate information is stored in the HDF5 root data group.

Land-Model-Constants_Data

Table A5 describes the data fields in the Land-Model-Constants_Data group stored in the lmc file collection. This group contains fields that specify static/time-invariant parameters (or constants) of the Catchment Land Surface Model (CLSM) and its associated L-band Microwave Radiative Transfer Model (MWRTM). Time and space coordinate information is stored in the HDF5 root data group.

Note: Due to the time-invariant nature of the file contents, the lmc file collection consists of only one granule per data product version (as identified by a distinct Science Version ID).

Data Field Definitions 

Table A6 lists all Level-4 soil moisture data fields and their definitions. All fields are two-dimensional and Float32 unless otherwise indicated in the description.

Two-Dimensional Arrays

All SPL4SM HDF5 data fields have /cell_lat /cell_lon shape. This shape is a two-dimensional array, where each data field represents a specific grid cell in the 9 km global cylindrical EASE-Grid 2.0 as specified by the cell_lat and cell_lon arrays. For example, the field surface_temp (234,789) represents the land surface temperature of the grid cell located at cell_lat (234,789) and cell_lon (234,789), where cell_row (234,789)=234 and cell_column (234,789)=789.

Fill/Gap Values 

SMAP data products employ fill and gap values to indicate when no valid data appear in a particular data field and ensure that data fields retain the correct shape. Gap values locate portions of a data stream that do not appear in the output data file. Yet because the SPL4SM data product is partially based on modeling, gaps are not expected to occur in the SPL4SM data stream. Note, however, that there might well be 3-hour intervals for which no SMAP data were assimilated. This situation would be reflected in the aup collection when the total number of assimilated observations for the time interval in question is zero.

Fill values appear in the SPL4SM data product over ocean and water surfaces or for variables that are not meaningful (such as snow temperatures in the absence of snow). Fill values are also used, for example, in the aup file collection for all grid cells for which SMAP observations were not assimilated. The latter may occur for any of the following circumstances:

• There was no SMAP overpass for the grid cell in question during the assimilation time window.
• The SMAP observations were not available due to quality control, missing science or engineering input data, or any other reason in the Level-1, -2, or -3 processing algorithms.
• The SMAP observations were rejected for assimilation due to quality control by the SPL4SM algorithm.

SMAP data products employ a specific set of data values to connote that a field is fill. The selected values that represent fill are dependent on the data type.

No valid value in the SPL4SM data product is equal to the values that represent fill. If any exceptions should exist in the future, the SPL4SM content will provide a means for users to discern between fields that contain fill and fields that contain genuine data values.

Notations 

Table A7 lists the acronyms and abbreviations used in this document.