*Article* **Assessment of Multi-Scale SMOS and SMAP Soil Moisture Products across the Iberian Peninsula**

**Gerard Portal 1,2,\*, Thomas Jagdhuber 3, Mercè Vall-llossera 1,2, Adriano Camps 1,2, Miriam Pablos 2,4, Dara Entekhabi <sup>5</sup> and Maria Piles <sup>6</sup>**


Received: 20 December 2019; Accepted: 6 February 2020; Published: 8 February 2020

**Abstract:** In the last decade, technological advances led to the launch of two satellite missions dedicated to measure the Earth's surface soil moisture (SSM): the ESA's Soil Moisture and Ocean Salinity (SMOS) launched in 2009, and the NASA's Soil Moisture Active Passive (SMAP) launched in 2015. The two satellites have an L-band microwave radiometer on-board to measure the Earth's surface emission. These measurements (brightness temperatures TB) are then used to generate global maps of SSM every three days with a spatial resolution of about 30–40 km and a target accuracy of 0.04 m3/m3. To meet local applications needs, different approaches have been proposed to spatially disaggregate SMOS and SMAP TB or their SSM products. They rely on synergies between multi-sensor observations and are built upon different physical assumptions. In this study, temporal and spatial characteristics of six operational SSM products derived from SMOS and SMAP are assessed in order to diagnose their distinct features, and the rationale behind them. The study is focused on the Iberian Peninsula and covers the period from April 2015 to December 2017. A temporal inter-comparison analysis is carried out using in situ SSM data from the Soil Moisture Measurements Station Network of the University of Salamanca (REMEDHUS) to evaluate the impact of the spatial scale of the different products (1, 3, 9, 25, and 36 km), and their correspondence in terms of temporal dynamics. A spatial analysis is conducted for the whole Iberian Peninsula with emphasis on the added-value that the enhanced resolution products provide based on the microwave-optical (SMOS/ERA5/MODIS) or the active–passive microwave (SMAP/Sentinel-1) sensor fusion. Our results show overall agreement among time series of the products regardless their spatial scale when compared to in situ measurements. Still, higher spatial resolutions would be needed to capture local features such as small irrigated areas that are not dominant at the 1-km pixel scale. The degree to which spatial features are resolved by the enhanced resolution products depend on the multi-sensor synergies employed (at TB or soil moisture level), and on the nature of the fine-scale information used. The largest disparities between these products occur in forested areas, which may be related to the reduced sensitivity of high-resolution active microwave and optical data to soil properties under dense vegetation.

**Keywords:** soil moisture; moisture variability; temporal dynamics; moisture patterns; spatial disaggregation; Soil Moisture Active Passive (SMAP); Soil Moisture and Ocean Salinity (SMOS); REMEDHUS

#### **1. Introduction**

Soil moisture (SM) is an essential climate variable (ECV) which plays a crucial role in the interplay between the Earth's land and atmospheric processes [1]. It is involved in the energy flux partition into latent and sensible heat from the land to the atmosphere. SM is closely linked to the soil evaporation, plant transpiration, and the allocation of precipitation into runoff, subsurface flow, and infiltration. Advancing our physical understanding of these land-atmosphere processes and interactions [2] is key for several climate and hydrological applications, such as drought and flood prediction, and weather and climate forecasting. Passive and active microwave sensors (radiometers and radars, respectively) are sensitive to the soil dielectric constant and allow estimation of surface soil moisture (SSM) [3]. Among microwave frequencies, measurements at L-band (1–2 GHz) have a higher soil penetration depth and are less affected by soil roughness, vegetation, and atmospheric effects than at higher frequencies (e.g., C- or X-bands) [4,5].

Currently, there are two L-band missions in orbit which were specifically devoted to measure SSM: (i) SMOS (Soil Moisture and Ocean Salinity) launched by the ESA (European Space Agency) in November 2009, and (ii) SMAP (Soil Moisture Active and Passive) launched by the NASA (National Aeronautics and Space Administration) in January 2015. Both systems have antennas with about a 6-meter aperture. The resulting brightness temperature measurements have about 40 km resolution using the half-power or −3 dB definition.

The spatial resolution of SMOS and SMAP brightness temperatures (TB) and derived SSM maps are in the order of tens of kilometers. However, to fulfill the needs of a growing number of applications, such as monitoring the evolution of insect pests [6], the prevention of wild fires [7,8], and the early detection of forest decline [9], among others, a higher spatial detail (<1 km) is required. To bridge this gap and improve the spatial resolution of the SSM maps, a variety of spatial enhancement or spatial (sub-pixel) disaggregation approaches have been proposed [10]. They generally differ in the ancillary information they use and the physical assumptions they rely on [11]. Consequently, the performance of these disaggregation algorithms depends mainly on the multi-sensor synergies employed and on the nature of the fine-scale information used which, in turn, may also depend on the season, climate, and land cover. This makes a direct comparison very challenging, since their performance is intrinsically linked to the method and rationale, and can also be time and region dependent.

This paper focuses on the in-depth analysis of SMAP and SMOS radiometer-only based products (SMAP at 9 and 36 km, SMOS at 25 km) and on their enhanced products which are now operational. They are based on two well-known satellite-based downscaling techniques: the active/passive microwave data fusion (SMAP/Sentinel-1 at 1 km and 3 km) [12], and the optical/thermal and microwave data fusion (SMOS/ERA5/MODIS at 1 km) [13,14].

The active/passive microwave data combination aims at obtaining an optimal blend of the high accuracy of passive sensors and the high spatial resolution of active sensors. Microwave radiometers have a high radiometric sensitivity (leading to soil moisture accuracies on the order of 0.04 m3/m3) and a high revisit time (three days), but coarse spatial resolution, typically 30–40 km. Therefore, microwave radars, especially Synthetic Aperture Radars (SARs) step in, as their spatial resolution is significantly higher, in the range of some meters. However, the backscatter commonly has a low temporal resolution (around one week) and may be significantly affected by soil roughness and the soil-covering vegetation canopy, which complicates the active-only soil moisture retrieval.

High-resolution maps can be obtained by combining information from the active and passive sensors. For this reason, some studies carried out before and after the SMAP launch, analyzed the covariation between passive and active microwave observations. This covariation is mostly driven by soil moisture dynamics, but also depends on changes on vegetation cover and soil roughness conditions [15–17], as occurs with the backscatter. When the SMAP radar failed, about 4-months after its launch, a method to disaggregate the L-band radiometer TB using the C-band Sentinel-1 radar backscatter was developed [12,18]. This approach, based on the active/passive covariation, is now the baseline to provide high-resolution SMAP SSM maps at 1 and 3 km [18]. However, the Sentinel-1 measurements are at C-band which have reduced sensitivity for moderate to dense vegetation coverage (up to ~3 kg/m2).

The optical/thermal and microwave fusion technique takes advantage of the high spatial resolution of optical and thermal remote sensing and on the inverse relationship between the land surface temperature (LST), and the vegetation status, which can be related to the soil moisture content [19]. Note that optical and thermal electromagnetic waves have the drawback of being masked by clouds, whereas microwaves can provide continuous monitoring regardless of atmospheric and illumination conditions. Here we use the latest version [13] of the optical/thermal and microwave algorithm firstly developed by Piles et al. [20,21]. It is an integrative model that holds at the coarse and fine spatial scales. Information of a vegetation index (Normalized Difference Vegetation Index, NDVI) from the optical and LST from the thermal bands of MODIS (moderate resolution imaging spectroradiometer instrument, MODIS) instrument, together with SMOS data, are used to obtain the model coefficients at low resolution. These coefficients are then applied to obtain the SSM fields at high resolution. Since the presence of clouds masking the MODIS LST information resulted in a loss of spatial coverage, a cloud free version of the algorithm [14] was developed in which MODIS LST was replaced with modelled ERA5 climate reanalysis skin temperature from the European Centre for Medium-Range Weather Forecast (ECMWF). Although the spatial resolution of the ERA5 LST is degraded with respect to MODIS LST (33 km vs. 1 km, respectively), the coverage increases dramatically; a comparison study carried out over Australia and Spain showed that the results were consistent for both versions of the algorithm [14]. This cloud-free version of the algorithm is now in operations at the Barcelona Expert Center (BEC) [22].

The aim of this paper is to analyze the temporal and the spatial characteristics of low-resolution (native) and high-resolution (disaggregated) SSM products provided by the SMAP and SMOS missions, with special emphasis on the most recently developed high-resolution ones. The temporal analysis has been carried out in the central part of the Duero basin, Spain, where the dynamics of SMAP and SMOS products at different spatial scales are compared against the data provided by the REMEDHUS in situ network, and their spatial representativeness as well as their correspondence is assessed. A comparison of spatial patterns has been conducted for the whole Iberian Peninsula, with focus on the analysis of their differences and distinct features, as well as on understanding the possible impact of the physical assumptions and multi-sensor synergies in the fine-scale estimates.

The SMAP and SMOS-derived SSM data products as well as the hydrological and climatic variables used in this study are presented in Section 2. Section 3 explains briefly the methodology followed to conduct the temporal and spatial analyses on the different products. The results of these comparisons are shown in Section 4. Section 5 discusses the possible reasons for the mismatch found among the different SSM products. Finally, Section 6 provides main conclusions and perspectives from this study.
