Search

Drought is a natural, erratic phenomenon that has a widespread and significant impact on socioeconomic and environmental development. The early monitoring and evaluation of drought through forecasting models would allow the articulation of early control and mitigation strategies, thus achieving an optimal development in planning and preparation for climate change. Therefore, this research developed a methodology for spatiotemporal analysis of drought patterns using automatic learning tools in the dry corridor of Central America. To this end, some specific milestones were defined. These include: (i) To assess temporal and spatial meteorological and agricultural droughts events, (ii) Identifying and validate results of the spatiotemporal events using computer vision techniques and finally (iii) Implementing machine learning drought forecasting models. ERA 5 monthly land average dataset was used as input for index estimation, spatiotemporal analysis and forecasting models. The frequency of drought events was calculated using standardized SPI and SPEI indices for accumulation periods of 1,3,6,9. However, 3,6 allowed a more realistic analysis of the seasonal change conditions in the hydrological regime of the area and the identification of the existing teleconnection between drought events and scale propagation. Regarding the spatiotemporal dynamics, 97 drought events of greater extension were identified, which are generally originated in countries such as Guatemala, Nicaragua, and El Salvador between seasonal periods not longer than 7 months. Additionally, the suitability of automatic learning models such as SVR, ANN and deep learning such as LSTM for index forecasting (r2=0.80) and drought dynamics in a temporal window of 1 to 6 months ahead was verified with considerable performance. The presented methodology provides an important basis for drought characterization and forecasting through the integration of spatiotemporal tracking models and machine learning techniques. Therefore, the methodological development can be adapted as an instrument for monitoring and forecasting, articulated to management and early mitigation policies. Finally, we suggest adapting variables related to the orographic context, relief, land use and land cover change, for instance, to improve the forecasting performance of the exposed forecasting models., La sequía es un fenómeno natural y errático que tiene un impacto generalizado y significativo en el desarrollo socioeconómico y medioambiental. El seguimiento y la evaluación temprana de la de la sequía mediante modelos de previsión permitiría articular estrategias tempranas de control y mitigación estrategias de control y mitigación tempranas, logrando así un desarrollo óptimo en la planificación y preparación para el cambio climático. Por lo tanto, esta investigación desarrolló una metodología para el análisis espacio-temporal de patrones de sequía utilizando herramientas de aprendizaje automático en el corredor seco de Centroamérica. Para ello Para ello, se definieron algunos hitos específicos. Estos incluyen: (i) Evaluar temporal y espacialmente meteorológicos y agrícolas de las sequías, (ii) Identificar y validar los resultados de los eventos espacio-temporales mediante técnicas de visión por ordenador y, por último, (iii) Implementar modelos de previsión de sequías de aprendizaje automático. aprendizaje automático de modelos de previsión de sequías. Se utilizó el conjunto de datos mensuales del ERA 5 como entrada para la estimación del índice, el análisis espacio-temporal y los modelos de previsión. y los modelos de previsión. La frecuencia de los eventos de sequía se calculó utilizando índices SPI y SPEI estandarizados para períodos de acumulación de 1,3,6,9. Sin embargo, el 3,6 permitió un análisis más realista de las condiciones de cambio estacional en el régimen hidrológico de la zona y la identificación de la teleconexión existente entre los eventos de sequía y la escala propagación. En cuanto a la dinámica espacio-temporal, se identificaron 97 eventos de sequía de mayor extensión se identificaron 97 eventos de sequía de mayor extensión, que generalmente se originan en países como Guatemala, Nicaragua y El Salvador entre períodos estacionales no mayores a 7 meses. Además, se ha comprobado la idoneidad de modelos de aprendizaje automático como SVR, RNA y aprendizaje profundo como LSTM para la previsión de índices (r2=0.80) y la dinámica de la sequía en una ventana temporal de 1 a 6 meses por delante fue verificado con un rendimiento considerable. La metodología presentada proporciona una base importante para la caracterización y de la sequía a través de la integración de modelos de seguimiento espacio-temporal y técnicas de aprendizaje automático. y técnicas de aprendizaje automático. Por lo tanto, el desarrollo metodológico puede ser adaptado como un instrumento de seguimiento y previsión, articulado a políticas de gestión y mitigación temprana. Por último, sugerimos la adaptación de variables relacionadas con el contexto orográfico, el relieve, el uso del suelo y el cambio de la cobertura del suelo cambio de la cubierta del suelo, por ejemplo, para mejorar el rendimiento de las previsiones de los modelos de previsión expuestos., Maestría, Magíster en Ingeniería Civil

One-pot cascade synthesis of L-homoserine catalysed by transaminase and pyruvate-dependent aldolase starting from pyruvate, formaldehyde and L-alanine (Scheme 1) was investigated in this work. The proof of concept was shown to work before by Hernandez et al. (2017). In this work two types of biocatalysts were used and kinetically characterized ; i.e. separate enzymes available as cell free extract (CFE). The focus was put on kinetic investigation of this complex reaction. Detailed kinetic analysis was done and kinetic parameters were estimated from the experimental data collected by using the initial reaction rate method. Based on the reaction scheme and the estimated apparent kinetic parameters the kinetic model was developed and was combined with mass balances in the reactor for separate reaction steps, as well as for the cascade reaction. Scheme 1 The reaction scheme of the synthesis of L-homoserine. The proposed model described the data well for the separate reaction steps ; i.e. aldol addition of pyruvate and formaldehyde catalyzed by aldolase and transamination of L-alanine by the product of the aldol addition. The kinetic model of the cascade reaction was used to find and propose the strategy for the improvement of product concentration and volume productivity. It was found that the best way to perform the reaction is in the fed-batch reactor with continuous feed of formaldehyde and two additions of L-alanine and pyruvate during the course of experiment up to their initial concentrations. This strategy enabled the production of L- homoserine of 672 mM and 76.23 g L-1 d-1 after 25 hours. The kinetic modeling was found to be crucial for understanding of the reaction system, and the obtained process metrics present a good starting point for potential industrial application. K. Hernandez, J. Bujons, J. Joglar, S. J. Charnock, P. Domínguez de María, W. Dieter Fessner, P. Clapés, ACS Catal., 2017, 7(3), 1707- 1711, doi: https://doi.org/10.1021/acscatal.6b03181. Project CARBAZYMES-This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 635595.

Biocatalysis offers advantages over chemical synthesis as enzyme-catalyzed reactions are often highly enantioselective and regioselective and can be carried out at ambient temperature and atmospheric pressure. Carbon-carbon bond formation is the corner stone reaction of organic synthesis which allows the creation of complex compounds from the simpler ones. Aldolases catalyze the reversible carbon-carbon bond formation by aldol addition of a nucleophilic donor to an electrophilic aldehyde acceptor. In this work the enzymatic aldol addition between propanal and formaldehyde was studied. The reaction product (3-hydroxy-2-methylpropanal) is an interesting precursor which can be used for the synthesis of Roche ester. Enzyme D-fructose-6-phosphate aldolase (FSA) is used as a catalyst in this reaction. Three different FSA variants that catalyze this reaction are compared. Kinetic parameters for FSA VarI, FSA VarII and FSA VarIII are estimated from the dependence of the initial reaction rate on the concentrations of each reaction component using the initial reaction rate method (at substrate conversions less than 10%). Furthermore, the comparison of these enzymes is made by running batch reactions under same conditions. The results show which FSA variant is the most appropriate for the use in this biocatalytic aldol addition reaction.

Carbon-carbon bond formation between propionaldehyde and formaldehyde using the thermostable D-fructose-6-phosphate aldolase as a green and enantioselective catalyst was investigated. The reaction kinetics can be described by the double substrate Michaelis- Menten kinetics with included substrate inhibitions by formaldehyde and propanal. Mathematical model for the reaction in the batch reactor included self-addition of propanal catalyzed by FSA also described by Michaelis- Menten kinetics. Experiment performed in batch mode shows that the mathematical model describes the experimental data well. Due to inhibition by both substrates high concentration of the product cannot be achieved. Also, high substrate concentration cannot be used due to their negative effect on enzyme stability. Based on the model for reaction in batch reactor, conditions for fed batch reactor were simulated in order to produce higher amount of aldol product. Fed batch mode suits this reaction system better than batch mode, and once optimized it will give the high product concentrations.

Carbon-carbon bond formation is the essence of organic synthesis and provides the foundation for generating more complex compounds from the simpler ones.[1] Aldolases catalyze the reversible carbon–carbon bond formation by aldol addition of a nucleophilic donor to an electrophilic aldehyde acceptor.[2] In this work aldol addition of formaldehyde and propanal catalyzed by D-fructose-6-phosphate aldolase (FSA) was investigated (Fig 1). This reaction has not been studied up to now. Figure 1. Reaction scheme of aldol addition between formaldehyde and propanal catalyzed by FSA. The main focus was put on the kinetic model development, finding the best reactor mode and process conditions to obtain the maximum concentration of aldol product and ensure the best use of enzyme. Thus, kinetic analysis of the reaction system was done and the reaction rate was described by the two substrate Michaelis-Menten kinetic model with included inhibitions by both substrates. It was found that propanal is very reactive at reaction conditions and causes formation of side- products with and without the presence of enzyme. Thus, the kinetics of side-reactions were also investigated to ensure good applicability of the mathematical model, and to try minimizing the formation of side-products. Enzyme operational stability was also experimentally evaluated. The model was validated in the batch reactor. By combining the kinetic model for the enzymatic aldol addition, as well as for side-reactions that occur, with mass balances in batch and fed- batch reactor, simulations were done to find the optimal conditions for this reaction system. [1] Li, C.-J. Chem. Rev. 2005, 105, 3095- 3165. [2] Clapés, P. ; Fessner, W.-D. ; Sprenger, G.A. ; Samland, A.K. Curr. Opin. Chem. Biol. 2010, 14, 154–167. Project CARBAZYMES-This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 635595.

In this work aldol addition of formaldehyde and propanal catalyzed by D-fructose-6-phosphate aldolase D6Q variant (FSA D6Q) was investigated. This reaction has not been studied up to now. The main focus was put on the kinetic model development, finding the best reactor mode and process conditions to obtain the maximum concentration of aldol product. Kinetic analysis of the reaction system was done and the reaction rate was described by the two substrate Michaelis-Menten kinetic model with included inhibitions by both substrates. It was found that propanal is very reactive at reaction conditions and causes formation of side-products with and without the presence of enzyme. Therefore, the kinetics of side-reactions were also investigated to ensure better applicability of the developed mathematical model, and to try minimizing the formation of side-products. The influence of pH value on the reaction is taken into account since a better product stability is achieved at pH 7 than at pH 8. The mathematical model was validated in the batch and fed-batch reaction mode. It is shown that the model describes the experimental data well. By using the simulation the conditions are calculated for performing a reaction in fed-batch mode that gives the highest product yield. Up to now the highest volume productivity of 287.3 g/L/d was obtained by this enzymatic synthesis.