Soil water content plays a critical role in agro-hydrology since it regulates the rainfall partition between surface runoff and infiltration and, the energy partition between sensible and latent heat fluxes. Current thermal inertia models characterize the spatial and temporal variability of water content by assuming a sinusoidal behavior of the land surface temperature between subsequent acquisitions. Such behavior implicitly supposes clear sky during the whole interval between the thermal acquisitions; but, since this assumption is not necessarily verified even if sky is clear at the exact epoch of acquisition, , the accuracy of the model may be questioned due to spatial and temporal variability of cloud coverage. During the irrigation season, cloud coverage exhibits a quite regular daily behavior, which, when rendered in probabilistic terms, allows for an a-priory evaluation of the most likely suitable pair of images to estimate thermal inertia, given the results of the satellite passes. In turn, the water content of soil is estimated through thermal inertia by coupling diurnal optical and nighttime thermal images, e.g. as acquired by MODIS sensor on board polar orbiting satellites AQUA and TERRA, which have spatial resolution high enough to cope with typical agricultural applications. The method relies on the availability of the shortwave albedo and, at least, two daily thermographs preferably acquired in specific epochs of the day: the first at sunset when latent and heat fluxes are negligible; the second just before sunrise, when surface soil temperature reaches its minimum. Unfortunately, high resolution thermal images are often not available in those specific epochs, so that the accuracy of estimate accuracy decays even severely. In this perspective the paper, following previous contributions by some of the authors of the present paper <sup>[1-4]</sup>, proposes exploiting SEVIRI data, characterized by higher acquisition rate but coarser spatial resolution as available from geostationary platform, to supplement MODIS data in a twofold way: i) by allowing to verify, by means of cloud detection algorithms, the hypothesis of clear sky throughout the time; ii) by synthesizing a high spatial/high temporal resolution sequence of images, through fusion of MODIS and SEVIRI data via Bayesian smoothing. A first validation of the latter method is achieved by comparing the results with in situ micro-meteorological measurements.
Soil water content is directly connected with soil evaporation and plant transpiration processes; in particular, soil water
content within the root zone, is readily available to evapotranspiration. Thus, in agricultural sciences, the assessment of
the spatial distribution of soil water content could be of utmost importance in evaluating crop water requirement.
In spite of limitations to applicability due to contingent cloud cover, water content of the upper part of the soil can be
determined by applying the thermal inertia approach by coupling optical and thermal infrared images. The thermal
inertia formulation, rigorously retrieved on bare soil, has been also verified on soils partially covered by vegetation. In
each case, one of the crucial steps is the assessment of the phase difference between surface temperature and solar
irradiation. Different approaches allow determining this latter parameter.
To this aim, three formulations to retrieve the phase difference were tested: i) the first, assuming a spatially constant
value based on the knowledge of the time when maximum surface temperature occurs; ii) other two methods, allowing
determining its spatial distribution through three or four thermographies.
In this framework, this research is focused to establish the simplest operational approach providing reliable results over
time using low-resolution MODIS images collected over an agricultural area of South Italy (Campania). Temporal
evolution of the remote sensing estimates have been compared to data collected by the micro-meteorological station
installed in a vineyard within the area.
Accurate estimation of physical quantities depends on the availability of High Resolution (HR) observations of
the Earth surface. However, due to the unavoidable tradeoff between spatial and time resolution, the acquisition
instants of HR data hardly coincides with those required by the estimation algorithms. A possible solution
consists in constructing a synthetic HR observation at a given time k by exploiting Low Resolution (LR) and
HR data acquired at different instants. In this work we recast this issue as a smoothing problem, thus focusing
on cases in which observations acquired both before and after time k are available. The proposed approach is
validated on a region of interest for the IRRISAT irrigation management project in which the surface thermal
inertia estimation, requiring multiple HR images at specific instants, constitute a key step.
In irrigation management the estimation of the radiometric surface temperature is of fundamental importance in evaluating the spatial distribution of land surface evapotranspiration. However, obtaining both high spatial and temporal resolutions data is impossible for any real sensor. In this paper we propose and investigate the use of sequential Bayesian techniques for integrating heterogeneous data with complementary features. A validation is performed by means of images acquired from SEVIRI and MODIS sensors in the thermal channels IR 10:8 and 31, respectively.
Information extraction from remotely sensed images acquired in the visible and near-infrared (VNIR) frequency
range strongly depends on an accurate cloud pixel screening. Indeed, many remote sensing applications require a
preliminary cloud detection phase to obtain profitable results. In this paper we propose to integrate the potential
of the MAP-MRF methodology with the multispectral approach for augmenting the capability of the algorithm
to detect cloudy pixels. In particular the proposed technique combines information from some SEVIRI sensor
channels (in particular the channels 0.64ìm, 1.6ìm, 3.9ìm, 7.3ìm and 10.8ìm) with the classification obtained
by the MAP-MRF method in the 0.8ìm channel in order to discriminate between snowy and cloudy pixels.
The validation is performed on challenging images of Alps mountains acquired by the SEVIRI sensor during
winter months. Results show significant improvements with respect to existing methods. In particular we
highlight a more precise classification at the cloud borders and a considerable reduction of unsolicited holes
inside the cloud masses.u