Field campaigns are a very important component of the process of validating satellite-based geophysical products. They can also strengthen collaboration within and across disciplines that can answer broader science questions. The primary driver for SMAPVEX16 is to address specific issues identified during the first year of SMAP post-launch validation. The top priority will be to contribute to improving the soil moisture algorithms.
Role of Field Campaigns in SMAP Cal/Val
The Soil Moisture Active Passive (SMAP) mission currently provides Level 2 to 4 soil moisture products derived from its L-band passive microwave radiometer. Due to aggressive efforts to mitigate radio frequency interference, the accuracy and temporal and spatial coverage of the passive-based soil moisture products have been very good (Chan et al. 2016). Following the loss of its radar that was to provide a higher resolution soil moisture product, SMAP is currently exploring a suite of approaches to enhance the spatial resolution of the soil moisture products. These include combining active observations from the European Space Agency (ESA) Sentinel 1 radars with the SMAP passive radiometer.
The SMAP post-launch Cal/Val activities are intended both to assess the quality of the mission products and to support analyses that lead to their improvement. A suite of complementary methodologies is being employed (Table 1) that will result in a robust global assessment. Beyond the in-orbit checkout (IOC) (launch + 3 months), the first year of operations focused on assessing the products. However, Cal/Val is an activity that continues throughout the mission life, and some anomalies and poor performance issues were identified that would require intensive efforts to resolve.
As noted in Table 1, field campaigns are intended to provide detailed information about targeted Cal/Val sites. It is recognized that only a limited number of field campaigns can be conducted due to their cost and the extensive logistics and human resources required (especially if airborne instruments are involved). It should also be noted that data latency is expected in these campaigns because they are typically one-of-a-kind efforts. In planning post-launch Cal/Val activities it is important that airborne instruments in field campaigns are linked with objectives that can only be addressed using airborne systems.
Science Objectives for a Post-Launch SMAP Aircraft-Based Field Campaign
Summarizing from above, post-launch Cal/Val of SMAP involves the assessment of products and improving the performance of the algorithms. The SMAP Project Science Team (ST) and the Cal/Val Working Group provide guidance to the Cal/Val activities of SMAP. Over the course of several meetings and workshops a consensus on priorities has been developed on what field campaigns can and should address. The following is a list of science issues that can only be addressed using an aircraft-based campaign
- Investigate and resolve anomalous observations and products
- Improve the up-scaling functions for CVS
- Contribute to a broader science/application objective
- Validate the L2_SM_AP algorithm process: pending new directions
- Understand the effects and contribution of heterogeneity on coarser resolution retrievals
- Evaluate the impact of known RFI sources on retrieval
- Analyze heterogeneity effects on L1 product calibration
For SMAPVEX16 the focus will be on the first item. It is also expected that it will contribute to the other items. Further details on some of the specific contributions follow.
Investigate and resolve anomalous observations and products
The beta-release assessment of the L2SMP (Chan et al. 2016) revealed that while the overall performance was quite good, there were some CVS with very large errors. These were the agricultural sites; South Fork (Iowa) and Carman (Manitoba). In addition, comparison of SMAP and SMOS global products indicated very different results over dense forest regions. Although SMAP produces a flagged retrieval for forest land cover, it is currently not a mission requirement. Therefore, for 2016 the priority will be to improve the retrieval at these agricultural sites. Efforts (and experiment plans) are being provided separately for the two sites. Approaches and protocols will be coordinated to the degree it is possible.
Improving up-scaling functions for CVS
SMAP utilizes a standard procedure for up-scaling the distributed in situ observations to the gridded product. This is based on Thiessen Polygons. There are several approaches that can be used to improve this process that include field campaigns involving more sampling points. SMAPVEX16 will incorporate this approach.
Contribution to a broader scientific/application objective
For SMAPVEX16-Iowa (IA), the field campaign will be coordinated with validation activities of two other NASA missions; the Orbital Carbon Observatory-2 (OCO-2) and the Global Precipitation Mission (GPM).
Validate the L2SMAP algorithm process: pending new directions
As noted above, SMAP is exploring new approaches to disaggregating the passive-based products to provide higher resolution soil moisture. SMAPVEX16 will be a testbed for these ideas. The SMAP project is currently implementing two products that utilize Sentinel 1a (and potentially 1b) C-band radar data. These will be 3 km and 9 km gridded products.
The SMAP Cal/Val program provides in situ, satellite, and model products that can be used to validate the soil moisture products. However, none of these resources can validate the entire L2SMAP retrieval process (disaggregation and retrieval) on a L2SMP basis. Using an aircraft-based radiometer to provide coverage of SMAP 3 and 9 km data cells is the only way to validate this critical component of the mission. Validating the disaggregation approach and the SM at 9 km will provide increased confidence in the product and/or insights that can be implemented to improve the algorithm.
The validation data sets generated in SMAPVEX16, in particular the maps of soil moisture provided by the aircraft can also be used to evaluate other approaches that can be used for disaggregation, such as those based on thermal and optical data.
SMAPVEX16 Aircraft Experiment Concept
The basic design calls for aircraft flights over the L2SMP validation grid cell at the CVS using a higher resolution SMAP simulator. A key requirement of the experiment is higher spatial resolution L-band brightness temperature (TB) coverage of the entire site that can be used to generate TB values for the L2SMAP (9 km and 3 km). The SMAP L2SMP algorithm(s) would be used with this data to generate SM products at both spatial resolutions using the ground sampling to validate the higher spatial resolution retrievals. A variety of soil moisture and vegetation conditions are needed for a robust analysis. Therefore, flights will be conducted over two IOPs at each of the two domains (Iowa-IA and Manitoba-MB). One IOP will be early in the growing season and the other during maximum biomass. Aircraft flights will be focused on the SMAP morning overpasses.
The dates of the IOPs during two intensive observing periods (IOPs) at each are listed below. These will be bridged with temporary stations and tower-based radiometer/radar observations.
Based upon the justification provided, an aircraft-based passive instrument is an essential requirement since it is the only means by which the process of brightness temperature disaggregation can be assessed. The aircraft based radiometer must be capable of providing coverage of an entire SMAP L2SMP grid cell within ~3 hours in order to minimize diurnal temperature and soil profile variability effects. Spatial resolution of 3 km or better is desirable if all objectives are to be satisfied. The instrument must be stable during flights and well calibrated.
The incidence angle should be centered at 40 degrees and the beam width should be as small as possible considering all other constraints. The antenna pattern should be well known.
For SMAPVEX16 the well-established Passive Active L-band System (PALS) will be utilized (Colliander et al., 2015). It was converted to a scanner that will be capable of meeting the experiment requirement. Details are provided in later sections.
Detailed Experiment Plans
Experiment plans and operations were developed by two teams with some overlap and coordination. The Iowa plan is provided here:
The Manitoba plan will be found here:
Chan, S., R. Bindlish, P. O’Neill, E. Njoku, T.J. Jackson, A. Colliander, F. Chen, M. Burgin, R.S. Dunbar, J. Piepmeier, S. Yueh, D. Entekhabi, M.H. Cosh, M.S. Seyfried, D.D. Bosch, P. Starks, D.C. Goodrich, J.H. Prueger, W.T. Crow, T. Caldwell, J. Walker, X. Wu, A. Pacheco, H. McNairn, M.C. Anderson. (2016). Assessment of the SMAP Level 2 passive Soil Moisture Product, IEEE Transactions on Geoscience and Remote Sensing, in press.
Colliander, A., T.J. Jackson, H. McNairn, S. Chazanoff, S. Dinardo, B. Latham, I. O'Dwyer, W. Chun, S. Yueh, E. Njoku. (2015). Comparison of Airborne Passive and Active L-Band System (PALS) Brightness Temperature Measurements to SMOS Observations During the SMAP Validation Experiment 2012 (SMAPVEX12). IEEE Transactions on Geoscience and Remote Sensing, Vol. 12, No. 4, April 2015.