Search

Winter snowmelt events are more significant in places where the snowpack and air temperatures are closer to the freezing point, however these conditions will become more important in other snow-dominated basins in a warmer climate. Winter warm spells (WWS) are extreme temperature anomalies that amplify snowmelt and sublimation in mountain regions with uncertain consequences to timing and volume of water resources. This study aims to understand the effect of WWS on snowpack ablation patterns in the mountainous western North America and how they might change under a warmer climate. We use two convection permitting regional climate model simulations to represent historical (2001-2013) and future atmospheric and surface conditions.Although days with ablation during WWS represent a small fraction (8.3%, 0.6 days on average), 55% of total ablation occurs during WWS over regions with significant snowpack (mean peak snow water equivalent over 150 mm). Consistently, a larger ablation rate (53%) is found during WWS than non-WWS events. Furthermore, ablation rates during WWS in humid regions are larger in a warmer climate than those of the dry regions, which is explained by the differences in the energy balance and the snowpack cold content. WWS predominantly drive snowmelt (93.8%) rather than sublimation (6.2%), which has relevance to water resources such as flood risk, soil moisture, and streamflow response. This study provides a comprehensive description of the impact of extreme temperature events and a warmer climate over our changing snowpack. We acknowledge financial support by CRHIAM Project ANID/FONDAP/15130015, and the Anillo project ACT-210080., The 28th IUGG General Assembly (IUGG2023) (Berlin 2023)

Worldwide landslides correspond to one of the most dangerous geological events due to their destructive power and unpredictable nature. In the Chilean Patagonia, the SERNAGEOMIN (Chilean Geological Survey Service) has detected 2533 landslide events in the Northern Patagonia (42. 7∘S, 72. 4∘W) alone, a small area compared to the whole Chilean Patagonia. However, only 11 evens have known date. Consequently, it is not possible to associate temporal triggers and mechanisms that control such events, resulting in a lack of understanding of the factors that enable landslides. This work aims to detect landslides and identify the main environmental variables (climatic and geomorphological) that explain their occurrence using machine learning methods. We will address the following research questions: 1) How can a temporal landslide dataset be built using Landsat images and Google Earth Engine in Northern Patagonia? 2) Once the landslides and their timing have been detected, what are the main variables that condition the landslide processes? In our work, we developed a temporal dataset of landslides for the northern Patagonia of Chile. We used three machine learning approaches, where it was possible to identify the main environmental variables that allow us to predict their generation. Statistical models show that during the last 19 years, there has been complex interaction between different environmental variables that have influenced the activity of landslides. Climatic indices, indicators of extreme events, have a high incidence in the events’ predictive capacity. However, the most important are those linked to Patagonia’s tectonic context. In particular, the time elapsed after the eruptive event of the Chaiten volcano. Finally, the Liquine-Ofqui Fault System’s presence extending throughout the entire north of Patagonia has generated discontinuities at a general level, causing significant geomorphological instability. The study area has a relief in evolution and reactive to climatic conditions. Therefore, we highlight the need to understand better the interaction between geological and climatic processes and the future impact of these natural hazards. We emphasize the importance of analyzing landslide controls, considering both geomorphological variables and sporadic geological events.

Fast and accurate diagnosis is critical for the triage and management of pneumonia, particularly in the current scenario of a COVID-19 pandemic, where this pathology is a major symptom of the infection. With the objective of providing tools for that purpose, this study assesses the potential of three textural image characterisation methods: radiomics, fractal dimension and the recently developed superpixel-based histon, as biomarkers to be used for training Artificial Intelligence (AI) models in order to detect pneumonia in chest X-ray images. Models generated from three different AI algorithms have been studied: K-Nearest Neighbors, Support Vector Machine and Random Forest. Two open-access image datasets were used in this study. In the first one, a dataset composed of paediatric chest X-ray, the best performing generated models achieved an 83.3% accuracy with 89% sensitivity for radiomics, 89.9% accuracy with 93.6% sensitivity for fractal dimension and 91.3% accuracy with 90.5% sensitivity for superpixels based histon. Second, a dataset derived from an image repository developed primarily as a tool for studying COVID-19 was used. For this dataset, the best performing generated models resulted in a 95.3% accuracy with 99.2% sensitivity for radiomics, 99% accuracy with 100% sensitivity for fractal dimension and 99% accuracy with 98.6% sensitivity for superpixel-based histons. The results confirm the validity of the tested methods as reliable and easy-to-implement automatic diagnostic tools for pneumonia.

Accurate and up-to-date information on the spatial and geographical characteristics of agricultural areas is an indispensable value for the various activities related to agriculture and research. Most agricultural studies and policies are carried out at the field level, for which precise boundaries are required. Today, high-resolution remote sensing images provide useful spatial information for plot delineation; however, manual processing is time-consuming and prone to human error. The objective of this paper is to explore the potential of deep learning (DL) approach, in particular a convolutional neural network (CNN) model, for the automatic outlining of agricultural plot boundaries from orthophotos over large areas with a heterogeneous landscape. Since DL approaches require a large amount of labeled data to learn, we have exploited the open data from the Land Parcel Identification System (LPIS) from the Chartered Community of Navarre, Spain. The boundaries of the agricultural plots obtained from our methodology were compared with those obtained using a state-of-the-art methodology known as gPb-UCM (global probability of boundary followed by ultrametric contour map) through an error measurement called the boundary displacement error index (BDE). In BDE terms, the results obtained by our method outperform those obtained from the gPb-UCM method. In this regard, CNN models trained with LPIS data are a useful and powerful tool that would reduce intensive manual labor in outlining agricultural plots.

Producción Científica, Wood microdensitometry provides an integrated measurement of inter and intra-annual changes in wood anatomy and lignification. Although it can be acquired through a wide array of techniques, X-ray-based techniques are still the standard. Conversion of a grayscale X-ray image to density and annual ring boundaries delimitation is performed through image analysis software. Proprietary software has dominated these applications, albeit Free Open Source Software (FOSS) has been developed recently. We present ρ-MtreeRing, a user-friendly FOSS that streamlines the entire microdensitometry analysis process through a graphical user interface based on Shiny R Software without any programming knowledge. We compared the results of this program with the most widely used commercial software (WinDendro), showing the validity of the results. ρ-MtreeRing can be personalized and developed by the microdensitometry research community., Junta de Castilla y León (VA171P20), Ministerio de Ciencia e Innovación y Ministerio de Universidades grant number CGL2017-87309-P (MGH PRE2018-084106) and project PROWARM (PID2020-118444GA-I00), Universidad Politécnica de Madrid (through Project RP200060107), Centro de Recursos Hídricos para la Agricultura y la Minería (ANID/FONDAP/15130015)

The continuous growth of the world’s population requires immediate action to ensure food security. Sorghum is among the five most-produced cereals and is a dietary staple in many developing countries. Therefore, it is of great importance to obtain precise information for improving cereal productivity. An indicator for estimating sorghum yields is the number of crop heads in different branching arrangements. Approaches based on image processing and artificial intelligence have proved useful for automatically and efficiently obtaining this type of information for different crops. However, their application to sorghum crops presents some additional challenges owing to differences in the shape and color of sorghum heads. In this study, a methodology to detect sorghum heads in unmanned aerial vehicle imagery was investigated, and its performance was evaluated using a standard quality index in object detection problems (mean average precision). Specifically, test-time-augmentation (TTA) techniques have been implemented using a set of geometrical and color transformations selected according to the sorghum plant imagery requiring analysis, as well as four different ensemble learning methods. Because these methods are weighted, two different approaches for calculating these weights to improve sorghum head detection have been proposed. The results show that in sorghum head detection, TTA strategies outperform detection based only on individual transformed testing sets. Moreover, these results were improved by the use of different weights during the ensemble of TTA results.

Efficient water management in agriculture requires an accurate estimation of evapotranspiration (ET). Even though local measurements can be used to estimate the components of the surface energy balance, these values cannot be extrapolated to large areas due to the heterogeneity and complexity of agricultural and natural land surfaces and the dynamic nature of their heat processes. This extrapolation can be done by using satellite imagery, which provides information in the infrared thermal band; however, this band is not available in most current operational remote sensors. Our work hypothesis is that it is possible to generate a spatially distributed estimation of $\text{ET}_{d}$ without thermal band by using nonparametric models as regression random forest models (RRFM). Six Landsat-7 scenes were used to generate the RRFM. Results were evaluated by comparing the values of $\text{ET}_{d}$ provided by RRFM with that obtained using the surface energy balance model. It has been shown that the results generated by RRFM present a good agreement with METRIC (mapping ET at high resolution using internalized calibration) results, both quantitatively and qualitatively, especially for agricultural vegetation and forest land covers. Moreover, it has been detected that the RRFM estimation quality depends on the meteorological conditions on the days previous to the satellite register. It can be concluded that the $\text{ET}_{d}$ estimated by the RRFM would be feasible for real applications when the thermal band is not available.

A correct delineation of agricultural parcels is a primary requirement for any parcel-based application such as the estimate of agricultural subsidies. Currently, high-resolution remote-sensing images provide useful spatial information to delineate parcels; however, their manual processing is highly time consuming. Thus, it is necessary to create methods which allow performing this task automatically. In this work, the use of a machine-learning algorithm to delineate agricultural parcels is explored through a novel methodology. The proposed methodology combines superpixels and supervised classification in order to determine which adjacent superpixels should be merged, transforming the segmentation issue into a machine learning matter. A visual evaluation of results obtained by the methodology applied to two areas of a high-resolution satellite image of fragmented agricultural landscape points out that the use of machine-learning algorithm for this task is promising.

Producción Científica, Efficient water management in agriculture requires a precise estimate of evapotranspiration ( ET ). Although local measurements can be used to estimate surface energy balance components, these values cannot be extrapolated to large areas due to the heterogeneity and complexity of agriculture environment. This extrapolation can be done using satellite images that provide information in visible and thermal infrared region of the electromagnetic spectrum; however, most current satellite sensors do not provide this end, but they do include a set of spectral bands that allow the radiometric behavior of vegetation that is highly correlated with the ET . In this context, our working hypothesis states that it is possible to generate a strategy of integration and harmonization of the Normalized Difference Vegetation Index ( NDVI ) obtained from Landsat-8 ( L8 ) and Sentinel-2 ( S2 ) sensors in order to obtain an NDVI time series used to estimate ET through fit equations specific to each crop type during an agricultural season (December 2017–March 2018). Based on the obtained results it was concluded that it is possible to estimate ET using an NDVI time series by integrating data from both sensors L8 and S2 , which allowed to carry out an updated seasonal water balance over study site, improving the irrigation water management both at plot and water distribution system scale., Fondo de Fomento al Desarrollo Científico y Tecnológico (project IT18I0008), Water Research Center For Agriculture and Mining (project CONICYT–FONDAP–15130015).

Xylem is a vascular tissue that conducts sap (water and dissolved minerals) from the roots to the rest of the plant while providing physical support and resources. Sap is conducted within dead hollow cells (called vessels in flowering plants) arranged to form long pipes. Once formed, vessels do not change their structure and last from years to millennia. Vessels’ configuration (size, abundance, and spatial pattern) constitutes a record of the plant–environment relationship, and therefore, a tool for monitoring responses at the plant and ecosystem level. This information can be extracted through quantitative anatomy; however, the effort to identify and measure hundreds of thousands of conductive cells is an inconvenience to the progress needed to have solid assessments of the anatomical–environment relationship. In this paper, we propose an automatic methodology based on convolutional neural networks to segment xylem vessels. It includes a post-processing stage based on the use of redundant information to improve the performance of the outcome and make it useful in different sample configurations. Three different neural networks were tested obtaining similar results (pixel accuracy about 90%), which indicates that the methodology can be effectively used for segmentation of xylem vessels into images with non-homogeneous variations of illumination. The development of accurate automatic tools using CNNs would reduce the entry barriers associated with quantitative xylem anatomy expanding the use of this technique by the scientific community.

The quantification of evaporation is very important for water resource management, determination of water availability and environmental modeling. Evaporation (E) is a key component of the hydrological cycle, which becomes more relevant under a changing climate scenario, where increases in temperature and E are projected on a regional scale. This study evaluated the hourly applicability of 21 methods to estimate E in a mountain lake. Models were grouped as combination, solar radiation-temperature and mass transfer methods. Estimated E by all models were compared to three measurement campaigns with an eddy covariance system in Lake Laja during the month of January (22 to 28), March (17 to 23) and November (09 to 15) of 2016. The average evaporated water during the measurement sessions were 3.40, 3.38 and 1.89 mm d&minus, 1, respectively. Best model performance was obtained with models whose main E principles are heat flow and water vapor flow. The best performance in this group was the Penman model using a calibrated wind function with a determination coefficient (R2) of 0.91, Nash&ndash, Sutcliffe coefficient (NS) of 0.9, and index of agreement (W) of 0.98. Comparing daytime and nighttime hours, it was found that the daytime E in the lake is best explained by the product of the vapor pressure deficit and wind speed, while the nighttime E was explained by the flow of heat in the water. The results highlight the importance of the analysis of diurnal dynamics of water flux and energy stored in water to better understand the E in water bodies.

In the scenario of current water scarcity caused by climate change and increasing water demand for food production, farmers must adapt their water management practices by shifting from supply-driven water management to demand-driven water management, considering trade-offs among quality, quantity and costs. Thus, agricultural practices must take full advantage of technology, research and development and adapt to local requirements. Nowadays, remote sensing is a useful tool for estimating crop water demand (evapotranspiration) as well as mapping their spatial and temporal variability. In this work, we present a new methodology that allows the user to audit (ex post) the irrigation strategies of a blueberry field in central Chile using a decision support system for irrigation decision called AquaSat® as the main tool. This tool combines satellite information with field data and provides spatially distributed information on crop water use for managing irrigation at a farm scale. The main contribution of this work is to detail a new approach for irrigation management through the comparison of volume of applied water, against evapotranspiration and potential demand. This procedure allows the user to audit current irrigation management and to determine the impacts on productivity. From our results, we can conclude that the applied water levels used at the farm during both seasons throughout of the irrigation sector were insufficient to reach the potential blueberries yield.

Xylem is a vascular tissue that conveys water and dissolved minerals from the roots to the rest of the plant and also provides physical support. The most important cells present in xylem are called vessels. These cells are arranged to form long pipes that carry water through the tree. The identification, counting and subsequent characterization of xylem vessels is essential for monitoring tree health and its relationship with climatic conditions. Although automatic and semi-automatic image processing tools are available to analyze the structure of xylem at the cellular level, they usually require the supervision of an expert to obtain optimal segmentation, making it a highly time-consuming process. To overcome this limitation, a Convolutional Neural Network model was used to process digital images of 23 branch sections in order to segment the xylem vessels. The obtained results were compared with other two classical methods, Otsu's thresholding method, and an active contour method known as Chan-Vese segmentation algorithm. The obtained results show the potential of convolutional neural networks to overcome aspects such as non-homogeneous illumination of images, where conventional methods tend to obtain unsatisfactory results.

Thermal infrared (TIR) imagery is normally acquired at coarser pixel resolution than that of shortwave sensors on the same satellite platform. TIR resolution is often not suitable for monitoring crop conditions of fragmented farming lands, e.g., the accurate estimates of evapotranspiration (ET) based on surface energy balance from remote sensing for irrigation water management. Consequently, thermal sharpening techniques have been developed to sharpen TIR imagery to a shortwave band pixel resolution. However, most methods concentrate on the visual effects of the thermal sharpened images, and they treat the pixels as independent samples without considering their spatial context, which can give rise to adverse effects such as artifacts. In this work, a new thermal sharpening method called TS2uRF is proposed. The potential of superpixels (SP) combined with regression random forest (RRF) have been used to augment the spatial resolution of the Landsat 8 TIR (100 m) imagery to their visible (VIS) spatial resolution (30 m). The SP has allowed the contextual information on the land cover to be integrated, and RRF has allowed the relationship between five spectral indices and TIR data to be integrated into a single model. The TIR sharpened images obtained using the TS2uRF were compared with images obtained using the TsHARP, one of the most classic thermal sharpening techniques, evaluating the root-mean-square error (RMSE) and structural similarity index (SSIM) for measuring image quality. In all of the cases evaluated, the RMSE and SSIM of the images sharpened using the TS2uRF method outperform those obtained using TsHARP. In particular, the TS2uRF method has an average error of 1.14 °C (RMSE) lower than TsHARP, regarding SSIM, TS2uRF outperforms TsHARP on average by 0.218 . From the visual comparison, it has been shown that the TS2uRF methodology avoids the artifacts that appear in the enhanced images using the TsHARP method.

Producción Científica, The outlining of agricultural land is an important task for obtaining primary information used to create agricultural policies, estimate subsidies and agricultural insurance, and update agricultural geographical databases, among others. Most of the automatic and semi-automatic methods used for outlining agricultural plots using remotely sensed imagery are based on image segmentation. However, these approaches are usually sensitive to intra-plot variability and depend on the selection of the correct parameters, resulting in a poor performance due to the variability in the shape, size, and texture of the agricultural landscapes. In this work, a new methodology based on consensus image segmentation for outlining agricultural plots is presented. The proposed methodology combines segmentation at different scales—carried out using a superpixel (SP) method—and different dates from the same growing season to obtain a single segmentation of the agricultural plots. A visual and numerical comparison of the results provided by the proposed methodology with field-based data (ground truth) shows that the use of segmentation consensus is promising for outlining agricultural plots in a semi-supervised manner., Agencia Estatal de Investigación (AEI) y Fondo Europeo de Desarrollo Regional (FEDER) a través del proyecto ARTeMISat-2: Procesamiento Avanzado de Datos de Teledetección para el Seguimiento y Gestión Sostenible de Recursos Marinos y Terrestres en Ecosistemas Vulnerables (project CTM2016-77733-R), Centro de Investigaciones Hídricas para la Agricultura y la Minería (project CONICYT–FONDAP–1513001), Junta de Castilla y León - Fondo Europeo de Desarrollo Regional (project VA026P17)

Estimation of the crop water requirement is critical in the optimization of the agricultural production process, due to that yield and costs are directly affected by this estimation. Nowadays, remote sensing is a useful tool for estimating Evapotranspiration (ET), since it is possible to map their spatial and temporal variability. ET models using satellite images have been developed in the last decades, using in most cases the surface energy balance which has generated good ET representation in different study sites. One of these models is METRIC (Mapping EvapoTranspiration at high Resolution using Internalized Calibration), which estimates ET using mainly data from Landsat 7 and 5 images, and a physical-empirical basis to solve the surface energy balance. The main drawback of the METRIC model is the low robustness in the selection of two parameters called anchor pixels. Even though the rules to select anchor pixels are standardized, the procedure requires a user to choose the area where these pixels will be selected. In this sense, ET estimation is highly sensible to this selection, producing important differences when different anchor pixels are selected. In this study, a machine learning method is implemented through the GEOBIA (Geographic Object Based Image Analysis) approach for the identification of anchor objects, changing the focus from the pixels to the objects. Image segmentation and classification processes are used for an adequate selection of anchor objects, considering spectral and contextual information. The main contribution of this work proves that it is not necessary to choose an area to select the anchor parameters, improving the numerical stability of the model METRIC and increasing the robustness of the ET estimation. Results were validated by comparing the original selection of anchor pixels, as well as in-situ ET estimation using data obtained from Surface Renewal Stations, in sugar beat crops.

Thermal infrared (T IR) imagery is normally acquired at coarser pixel resolution than that of shortwave sensors on the same satellite platform. Often, T IR resolution is not suitable for monitoring crop conditions of individual fields or the impacts of land cover changes that are at significantly finer spatial scales. Consequently, thermal sharpening techniques have been developed to sharpen T IR imagery to shortwave band pixel resolutions. One of the most classic thermal sharpening technique is T sHARP . It uses a relationship between land surface temperature and normalized vegetation index (N DV I). However, there are several studies that prove that a single relationship between T IR and N DV I may only exist for a limited class of landscape. Our work hypothesis stated that it is possible to improve the spatial resolution of T IR imagery considering a relationship between vegetation and several soil spectral indexes and T IR as well the spatial context information. In this work, the potential of Superpixels (SP ) combined with Regression Random Forest (RRF ) is used to augmenting the spatial resolution of the Landsat 8 T IR (Band 10 and 11) imagery to their visible (V IS) spatial resolution. The SP allows to consider the contextual information over the land cover, and RF allows to integrate in a unique model the relationship between five spectral indices and T IR data. The results obtained by SP-RRF approach shows the potential of this methodology, compared with classical T sHARP method.

Efficient water management in agriculture requires an accurate estimation of evapotranspiration (ET). There are available several balance energy surface models that provide a daily ET estimation (ETd) spatially and temporarily distributed for different crops over wide areas. These models need infrared thermal spectral band (gathered from remotely sensors) to estimate sensible heat flux from the surface temperature. However, this spectral band is not available for most current operational remote sensors. Even though the good results provided by machine learning (ML) methods in many different areas, few works have applied these approaches for forecasting distributed ETd on space and time when aforementioned information is missing. However, these methods do not exploit the land surface characteristics and the relationships among land covers producing estimation errors. In this work, we have developed and evaluated a methodology that provides spatial distributed estimates of ETd without thermal information by means of Convolutional Neural Networks.

Recently, there has been a noteworthy increment of using images registered by unmanned aerial vehicles (UAV) in different remote sensing applications. Sensors boarded on UAVs has lower operational costs and complexity than other remote sensing platforms, quicker turnaround times as well as higher spatial resolution. Concerning this last aspect, particular attention has to be paid on the limitations of classical algorithms based on pixels when they are applied to high resolution images. The objective of this study is to investigate the capability of an OBIA methodology developed for the automatic generation of a digital terrain model of an agricultural area from Digital Elevation Model (DEM) and multispectral images registered by a Parrot Sequoia multispectral sensor board on a eBee SQ agricultural drone. The proposed methodology uses a superpixel approach for obtaining context and elevation information used for merging superpixels and at the same time eliminating objects such as trees in order to generate a Digital Terrain Model (DTM) of the analyzed area. Obtained results show the potential of the approach, in terms of accuracy, when it is compared with a DTM generated by manually eliminating objects.

The delineation of agricultural lands is an important task for obtaining primary information in order to establish agricultural policies. Most of the automatic and semiautomatic methods for delineating agricultural plots in remote sensing area are based on image segmentation. However, they usually are sensitive to intra-plot variability and depend on a correct parameter selection, resulting in a bad performance due to the variability of the agricultural landscapes. In this paper, a novel methodology for delineating agricultural plots is proposed. The methodology is based on the consensus of diverse superpixel segmentations at different scales. The join of different segmentations is intended to reduce the aforementioned problems. Proposed methodology has been applied to two regions of a Worldview-2 image of a fragmented agricultural landscape. A visual and numerical evaluation of the results indicates that the use of the consensus of segmentations is promising for delineating agricultural plots.

Land cover maps at different resolutions and mapping extents contribute to modeling and support decision making processes. Because land cover affects and is affected by climate change, it is listed among the 13 terrestrial essential climate variables. This paper describes the generation of a land cover map for Latin America and the Caribbean (LAC) for the year 2008. It was developed in the framework of the project Latin American Network for Monitoring and Studying of Natural Resources (SERENA), which has been developed within the GOFC-GOLD Latin American network of remote sensing and forest fires (RedLaTIF). The SERENA land cover map for LAC integrates: 1) the local expertise of SERENA network members to generate the training and validation data, 2) a methodology for land cover mapping based on decision trees using MODIS time series, and 3) class membership estimates to account for pixel heterogeneity issues. The discrete SERENA land cover product, derived from class memberships, yields an overall accuracy of 84% and includes an additional layer representing the estimated per-pixel confidence. The study demonstrates in detail the use of class memberships to better estimate the area of scarce classes with a scattered spatial distribution. The land cover map is already available as a printed wall map and will be released in digital format in the near future. The SERENA land cover map was produced with a legend and classification strategy similar to that used by the North American Land Change Monitoring System (NALCMS) to generate a land cover map of the North American continent, that will allow to combine both maps to generate consistent data across America facilitating continental monitoring and modeling. Fil: Blanco, Paula Daniela. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Nacional Patagónico; Argentina Fil: Colditz, Rene R.. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad; México Fil: López Saldaña, Gerardo. Universidade Técnica de Lisboa; Portugal Fil: Hardtke, Leonardo Andrés. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Nacional Patagónico; Argentina Fil: Llamas, Ricardo M.. Comisión Nacional para el Conocimiento y Uso de la Biodiversidad; México Fil: Mari, Nicolás A.. Instituto Nacional de Tecnología Agropecuaria. Centro Nacional de Investigaciones Agropecuarias. Centro de Investigaciones de Recursos Naturales. Instituto de Clima y Agua; Argentina Fil: Fischer, Angeles. Instituto Nacional de Tecnología Agropecuaria. Centro Nacional de Investigaciones Agropecuarias. Centro de Investigaciones de Recursos Naturales. Instituto de Clima y Agua; Argentina Fil: Caride, Constanza. Instituto Nacional de Tecnología Agropecuaria. Centro Nacional de Investigaciones Agropecuarias. Centro de Investigaciones de Recursos Naturales. Instituto de Clima y Agua; Argentina Fil: Aceñolaza, Pablo Gilberto. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro de Investigaciones Científicas y Transferencia de Tecnología a la Producción; Argentina. Universidad Autónoma de Entre Ríos; Argentina Fil: del Valle, Hector Francisco. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Nacional Patagónico; Argentina Fil: Lillo Saavedra, Mario. Universidad de Concepción; Chile Fil: Coronato, Fernando Raul. Consejo Nacional de Investigaciones Científicas y Técnicas. Centro Nacional Patagónico; Argentina Fil: Opazo, Sergio A.. Universidad de Magallanes; Chile Fil: Morelli, Fabiano. Centro de Previsao de Tempo e Estudos Climaticos. Instituto Nacional de Pesquisas Espaciais; Brasil Fil: Anaya, Jesús A.. Universidad de Medellin; Colombia Fil: Sione, Walter Fabian. Universidad Autónoma de Entre Ríos; Argentina. Universidad Nacional de Lujan; Argentina Fil: Zamboni, Pamela. Universidad Autónoma de Entre Ríos; Argentina Fil: Barrena Arroyo, Victor. Universidad Nacional Agraria La Molina; Perú

Precipitation is a key parameter for the study of climate change and variability and the detection and monitoring of natural disaster such as drought. Precipitation datasets that accurately capture the amount and spatial variability of rainfall is critical for drought monitoring and a wide range of other climate applications. This is challenging in many parts of the world, which often have a limited number of weather stations and/or historical data records. Satellite-derived precipitation products offer a viable alternative with several remotely sensed precipitation datasets now available with long historical data records (+30 years), which include the Climate Hazards Group InfraRed Precipitation with Station (CHIRPS) and Precipitation Estimation from Remotely Sensed Information using Artificial Neural Networks-Climate Data Record (PERSIANN-CDR) datasets. This study presents a comparative analysis of three historical satellite-based precipitation datasets that include Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA) 3B43 version 7 (1998-2015), PERSIANN-CDR (1983-2015) and CHIRPS 2.0 (1981-2015) over Chile to assess their performance across the country and evaluate their applicability for agricultural drought evaluation when used in the calculation of commonly used drought indicator as the Standardized Precipitation Index (SPI). In this analysis, 278 weather stations of in-situ rainfall measurements across Chile were initially compared to the satellite-based precipitation estimates. The study area (Chile) was divided into five latitudinal zones: North, North-Central, Central, South-Central and South to determine if there were a regional difference among these satellite-based estimates. Nine statistics were used to evaluate the performance of satellite products to estimate the amount and spatial distribution of historical rainfall across Chile. Hierarchical cluster analysis, k-means and singular value decomposition were used to analyze these datasets to better understand their similarities and differences in characterizing rainfall patterns across Chile. Monthly analysis showed that all satellite products highly overestimated precipitation in the arid North zone. However, there were no major difference between all three products from North to South-Central zones. Though, in the South zone, PERSIANN-CDR shows the lowest fit with high underestimation, further CHIRPS 2.0 and TMPA 3B43 v7 had better agreement with in-situ measurements. The accuracy of satellite products were highly dependent on the amount of monthly rainfall with the best results found during winter seasons and in zones (Central to South) with higher amounts of precipitation. PERSIANN-CDR and CHIRPS 2.0 were used to derive SPI at time-scale of 1, 3 and 6 months, both satellite products presented similar results when it was compared in-situ against satellite SPI's. Because of its higher spatial resolution that allows better characterizing of spatial variation in precipitation pattern, the CHIRPS 2.0 was used to mapping the SPI-3 over Chile. The results of this study show that in order to use the CHIRPS 2.0 and PERSIANN-CDR data sets in Chile to monitor spatial patterns in the rainfall and drought intensity conditions, these products should be calibrated to adjust for the overestimation/underestimation of precipitation geographically specially in the North zone and seasonally during the summer and spring months in the other zones.

The generation of labeled data for training automated methods used in the analysis of remote sensing images is a challenging task. Approaches as Active Learning aim to perform accurate classifications in a scenario of a few annotated data. These approaches generally require the interaction between the user and the machine learning method during training phase. However, in the remote sensing area, it is difficult to find a tool that facilitates this interaction. In this work, an interactive web-based platform to perform the training of method for remote sensing image annotation by means of an active learning approach is proposed. The platform integrates open-source GIS technologies and object-based approach in order to facilitate the interaction between the user and the active learning approach. While some usability aspects should be improved, the obtained results show the potential of the proposed platform.