Sub-seasonal drought forecasting is crucial for early warning in estimating agricultural production and optimizing irrigation management, as forecasting skills are relatively weak during this period. Soil moisture exhibits stronger persistence compared to other climate system quantities, which makes it especially influential in shaping land-atmosphere feedback, thus supplying a unique insight into drought forecasting. Relying on the soil moisture memory, this study investigates the combination of multiple deep-learning modules for sub-seasonal drought indices hindcast in the Huai River basin of China, using long-term ERA5-Land soil moisture records with a noise-assisted data analysis tool. The inter-compared deep-learning models include a hybrid model and a committee machine framework. The results show that the performance of the committee machine framework can be improved with the help of series decomposition and the forecasting skill is not impaired with the lead time increases. Overall, this study highlights the potential of combining deep-learning models with soil moisture memory analysis to improve sub-seasonal drought forecasting.

@en

The typical drivers of drought events are lower than normal precipitation and/or higher than normal evaporation. The region's characteristics may enhance or alleviate the severity of these events. Evaluating the combined effect of the multiple factors influencing droughts requires innovative approaches. This study applies hydrological modelling and a machine learning tool to assess the relationship between hydroclimatic characteristics and the severity of agricultural and hydrological droughts. The Soil Water Assessment Tool (SWAT) is used for hydrological modelling. Model outputs, soil moisture and streamflow, are used to calculate two drought indices, namely the Soil Moisture Deficit Index and the Standardized Streamflow Index. Then, drought indices are utilised to identify the agricultural and hydrological drought events during the analysis period, and the index categories are employed to describe their severity. Finally, the multivariate regression tree technique is applied to assess the relationship between hydroclimatic characteristics and the severity of agricultural and hydrological droughts.

Our research indicates that multiple parameters influence the severity of agricultural and hydrological droughts in the Cesar River basin. The upper part of the river valley is very susceptible to agricultural and hydrological drought. Precipitation shortfalls and high potential evapotranspiration drive severe agricultural drought, whereas limited precipitation influences severe hydrological drought. In the middle part of the river, inadequate rainfall partitioning and an unbalanced water cycle that favours water loss through evapotranspiration and limits percolation cause severe agricultural and hydrological drought conditions. Finally, droughts are moderate in the basin's southern part (Zapatosa marsh and the Serranía del Perijá foothills). Moderate sensitivity to agricultural and hydrological droughts is related to the capacity of the subbasins to retain water, which lowers evapotranspiration losses and promotes percolation. Results show that the presented methodology, combining hydrological modelling and a machine learning tool, provides valuable information about the interplay between the hydroclimatic factors that influence drought severity in the Cesar River basin.@en

Water utilities are urged to decrease their real water losses, not only to reduce costs but also to assure long-term sustainability. Hardware- and software-based techniques have been broadly used to locate leaks; within the latter, previous works that have used data-driven models mostly focused on single leaks. This paper presents a methodology to locate multiple leaks in water distribution networks employing pressure residuals. It consists of two phases: one is to produce training data for the data-driven model and cluster the nodes based on their leak-flow-rate-independent signatures using an adapted hierarchical agglomerative algorithm; the second is to locate the leaks using a top-down approach. To identify the leaking clusters and nodes, we employed a custom-built k-nearest neighbor (k-NN) algorithm that compares the test instances with the generated training data. This instance-to-instance comparison requires substantial computational resources for classification, which was overcome by the use of high-performance computing. The methodology was applied to a real network located in a European town, comprising 144 nodes and a total length of pipes of 24 km. Although its multiple inlets add redundancy to the network increasing the challenge of leak location, the method proved to obtain acceptable results to guide the field pinpointing activities. Nearly 70% of the areas determined by the clusters were identified with an accuracy of over 90% for leak flows above 3.0 L/s, and the leaking nodes were accurately detected over 50% of the time for leak flows above 4.0 L/s.

@en

Extreme peak runoff forecasting is still a challenge in hydrology. In fact, the use of traditional physically-based models is limited by the lack of sufficient data and the complexity of the inner hydrological processes. Here, we employ a Machine Learning technique, the Random Forest (RF) together with a combination of Feature Engineering (FE) strategies for adding physical knowledge to RF models and improving their forecasting performances. The FE strategies include precipitation-event classification according to hydrometeorological criteria and separation of flows into baseflow and directflow. We used ∼ 3.5 years of hourly precipitation information retrieved from two near-real-time satellite precipitation databases (PERSIANN-CCS and IMERG-ER), and runoff data at the outlet of a 3391-km² basin located in the tropical Andes of Ecuador. The developed models obtained Nash-Sutcliffe efficiencies varying from 0.86 to 0.59 for lead times between 1 and 6 h. The best performances were obtained for peak runoffs triggered by short-extension precipitation events (<50 km²) where infiltration- or saturation-excess runoff responses are well learned by the RF models. Conversely, the forecasting difficulty is associated with extensive precipitation events. For such conditions, a deeper characterization of the biophysical characteristics of the basin is encouraged for capturing the dynamic of directflow across multiple runoff responses. All in all, the potential to employ near-real-time satellite precipitation and the use of FE strategies for improving RF forecasting provides hydrologists with new tools for real-time runoff forecasting in remote or complex regions.

@en

This study explores the ways to introduce physical interpretability into the process of optimizing operating rules for multireservoir systems with multiple objectives. Prior studies applied the concept of direct policy search (DPS), in which the release policy is expressed as a set of parameterized functions (e.g., neural networks) that are optimized by simulating the performance of different parameter value combinations over a testing period. The problem with this approach is that the operators generally avoid adopting such artificial black-box functions for the direct real-time control of their systems, preferring simpler tools with a clear connection to the system's physics. This study addresses this mismatch by replacing the black-box functions in DPS with physically based parameterized operating rules, for example by directly using target levels in dams as decision variables. This leads to results that are physically interpretable and may be more acceptable to operators. The methodology proposed in this work is applied to a network of five reservoirs and four power plants in the Nechí catchment in Colombia, with four interests involved: average energy generation, firm energy generation, flood hazard, and flow regime alteration. The release policy is expressed depending on only 12 parameters, which significantly reduces the computational complexity compared to existing approaches of multiobjective DPS. The resulting four-dimensional Pareto-approximate set offers a variety of operational strategies from which operators may choose one that corresponds best to their preferences. For demonstration purposes, one particular optimized policy is selected and its parameter values are analyzed to illustrate how the physically based operating rules can be directly interpreted by the operators.

@en

Advances in near real-time rainstorm prediction using remote sensing have offered important opportunities for effective disaster management. However, this information is subject to several sources of systematic errors that need to be corrected. Temporal and spatial characteristics of both satellite and in-situ data can be combined to enhance the quality of storm estimates. In this study, we present a spatiotemporal object-based method to bias correct two sources of systematic error in satellites: displacement and volume. The method, Spatiotemporal Contiguous Object-based Rainfall Analysis for Bias Correction (ST-CORAbico), uses the spatiotemporal rainfall analysis ST-CORA incorporated with a multivariate kernel density storm segmentation for describing the main storm event characteristics (duration, spatial extension, volume, maximum intensity, centroid). Displacement and volume are corrected by adjusting the spatiotemporal structure and the intensity distribution, respectively. ST-CORAbico was applied to correct the early version of the Integrated Multi-satellite Retrievals for the Global Precipitation Mission (GPM-IMERG) over the Lower Mekong basin in Thailand during the monsoon season from 2014 to 2017. The performance of ST-CORABico is compared against the Distribution Transformation (DT) and Gamma Quantile Mapping (GQM) probabilistic methods. A total of 120 storm events identified over the study area were classified into short and long-lived storms by using a k-means cluster analysis method. Examples for both storm event types describe the error reduction due to location and magnitude by ST-CORAbico. The results showed that the displacement and magnitude correction made by ST-CORAbico considerably reduced RMSE and bias of GPM-IMERG. In both storm event types, this method showed a lower impact on the spatial correlation of the storm event. In comparison with DT and GQM, ST-CORAbico showed a superior performance, outperforming both approaches. This spatiotemporal bias correction method offers a new approach to enhance the accuracy of satellite-derived information for near real-time estimation of storm events.

@en

Lake Riogrande II, located in the central region of the Colombian Andes, has eutrophication due to a progressive increase of runoff pollution from upstream intensive dairy cattle and agricultural activities in the watershed. Public and private entities have invested in programs aimed at adopting the best agricultural management practices (BMP). BMP is a general formulation of recommended criteria and, therefore, not universally fit all problems. For example, the corrective measures taken so far demonstrate difficulties in selecting and allocating BMPs aligned to the space-time hydrological variability within the watershed. This research analyses the spatio-temporal dynamics of BMP pollution patterns. The study is built on the concept of critical source areas of runoff pollution from agricultural practices in the Colombia Andes. Fieldwork was conducted with the participation of farmers to collect spatial data of the current management operations for potato (Solanum tuberosum), tree tomato (Solanum betaceum), and dairy agriculture. A Soil Water Assessment model is used to simulate agricultural and hydrological processes. The model was calibrated using observed discharge, nitrate-N (NO₃ ^—N), and soluble phosphorus (P) concentrations at a monthly scale. The origin of pollution at the catchment scale was used applying the critical source areas (CSAs) method. This paper proposes for this analysis a new spatio-temporal CSA index (ST-CSA) to represent the behavior of the CSAs simultaneously in space and time. For this, several aggregated CSA in time were analyzed for monthly and annual time steps. Results indicate that there is an essential difference in the aggregation of CSA in time, and pollution location changes significantly using the ST-CSA. As expected, the highest number of CSAs occurs during the rainy months. However, these CSAs are located in the region with the lowest precipitation levels in the agricultural areas dominated by potato, tree tomato, and dairy agriculture located within the watershed. These areas vary significantly from 24.07% of the entire area of the basin (1034.348 km²), and it could reach a maximum value of 61.78%. Despite the model uncertainties, the results highlight the importance of identifying spatio-temporal CSAs to select BMPs with the highest potential of nitrogen and phosphorus losses reduction - such as the adaptation of fertilization schedules - applicable to the study watershed.

@en

In the last decades, advancements in computational science have greatly expanded the use of artificial neural networks (ANNs) in hydrogeology, including applications on groundwater forecast, variable selection, extended lead-times, and regime-specific analysis. However, ANN-model performance often omits the sensitivity to observational uncertainties in hydroclimate forcings. The goal of this paper is to implement a data-driven modeling framework for assessing the sensitivity of ANN-based groundwater forecasts to the uncertainties in observational inputs across space, time, and hydrological regimes. The objectives are two-folded. The first objective is to couple an ANN model with the PAWN sensitivity analysis (SA). The second objective is to evaluate the scale- and process-dependent sensitivities of groundwater forecasts to hydroclimate inputs, computing the sensitivity index in groundwater wells (1) across the whole time-series (for the global sensitivity analysis); (2) across the output sub-regions with conditions of water deficit and water surplus (for the ‘regional’ sensitivity analysis); and (3) at each time step (for the time-varying sensitivity analysis). The implementation of the ANN-PAWN occurs in 68 wells across the Northern High Plains aquifer, USA, with pre-time-step rainfall, evapotranspiration, snowmelt, streamflow, and groundwater measurements as inputs. Results show that evapotranspiration and rainfall are the major sources of uncertainty, with the latter being particularly relevant in water surplus conditions and the former in water deficit conditions. The time-varying sensitivity analysis leads to the identification of localized sensitivities to other sources of uncertainty, as snowmelt in spring or river flow during the annual peak period at the groundwater level.

@en

Understanding, characterizing, and predicting drought is vital for the reduction of its consequences. In the last few decades, many studies have moved drought analysis from the conventional lumped approach to a more spatiotemporal analysis. Two main developments have motivated this: one is global data availability and the other is the number of models developed to understand and quantify drought. The first one relates to information available from reanalysis products, and the second regards global and regional, distributed and semidistributed model data. Moreover, nowadays, different organizations provide drought monitoring information in near real time. However, a few spatiotemporal analysis studies have been developed slowly and the availability of comprehensive tools is still limited. This chapter proposes a new toolbox that performs the Spatio-Temporal ANalysis of Drought (STAND) in MATLAB, step by step. The toolbox collects some of the applications of previous studies and innovates new concepts on the characterization of drought. The methodologies here allow estimation of drought duration, severity (magnitude), and area, redefining the drought event in space and time. A key component in the analysis is the visualization of outcomes, as well as the spatial interpolation of pointwise data. The proposed STAND toolbox is explained and its use is illustrated with two large-scale examples (India and Mexico). The results have been compared with local reported information. STAND outcomes have been shown to help follow space–time events in terms of patterns, and provide information related to the characterization of extremes for drought analysis.@en

Drought indicators are of critical importance in characterization and forecasting. The use of the Standardized Precipitation Index (SPI) has increasingly become the main tool for drought analysis; however, the index lacks hydrological information useful as a proxy for other types of droughts. This study aims at evaluating the SPI against the Standardized Precipitation Evapotranspiration Index (SPEI) in the South Central Region of Vietnam. The indices were calculated using monthly rainfall and temperature data measurements from 30 rainfall and 13 temperature stations, during the period from 1977 to 2014. The study focuses on the spatial-temporal variations of drought events and therefore an area of 1680 grid cells of 4x4 km was selected. Inverse distance weighting was used to interpolate grid rainfall and temperature prior to drought indices estimating at multiple time scales (3, 6, 9, and 12 months). Drought severity was classified from gridded SPIs and SPEIs using a Non-Contiguous Drought Area (NCDA) approach. The result indicated that drought characteristics using the SPEI and NCDA can capture better historical drought conditions than that using the SPI and NCDA. This suggests an important role of temperature factor in the degree of drought severity. The analysis of spatial-temporal drought on the SPEI showed that the occurrence of moderate droughts in the study area was 1-2 years, and the highest percentage of drought in the area was observed in the summer-autumn season at all SPEI time scales. The results of this study may extend our understanding of natural drought mechanisms.

@en

Extreme hydrological events (EHEs), such as droughts and floods, vary spatially and temporally in nature. The increase in the number of events in the last few decades has motivated the research of the spatiotemporal variability of the future extreme precipitation and temperature. To study the consequences on the EHEs due to the uncertainty of projected climate changes, the analysis in more detail of precipitation and temperature, in space and time, is vital. In addition, for proper planning and decision-making process to address EHEs, understanding such climate changes requires more information. In this chapter we present a summarized assessment of the spatiotemporal variations of climate projections. A simplified way to aggregate global data is used for the spatiotemporal analysis of precipitation and temperature. To carry out this analysis, the Spatio-TempoRal distribution and Interannual VarIability of projections (STRIVIng) toolbox is proposed for statistical exploratory analysis of climate projections. Three large-scale applications were carried out for illustration: Dominican Republic (48,670 km²), Mexico (1,972,550 km²), and Amazon basin (6,171,148.7 km²). The methodology and toolbox presented here allow regions to be identified where the changes are expected to be more severe on precipitation and temperature, as well as months in which those changes are likely to occur. The STRIVIng toolbox is open source and helps to provide basic information to increase the interpretations and research in the space–time analysis of extremes.@en

The aim of this paper is to improve semiseasonal forecast of groundwater availability in response to climate variables, surface water availability, groundwater level variations, and human water management using a two-step data-driven modeling approach. First, we implement an ensemble of artificial neural networks (ANNs) for the 300 wells across the High Plains aquifer (USA). The modeling framework includes a method to choose the most relevant input variables and time lags; an assessment of the effect of exogenous variables on the predictive capabilities of models; and the estimation of the forecast skill based on the Nash-Sutcliffe efficiency (NSE) index, the normalized root mean square error, and the coefficient of determination (R²). Then, for the ANNs with low- accuracy, a MultiModel Combination (MuMoC) based on a hybrid of ANN and an instance-based learning method is applied. MuMoC uses forecasts from neighboring wells to improve the accuracy of ANNs. An exhaustive-search optimization algorithm is employed to select the best neighboring wells based on the cross correlation and predictive accuracy criteria. The results show high average ANN forecasting skills across the aquifer (average NSE > 0.9). Spatially distributed metrics of performance showed also higher error in areas of strong interaction between hydrometeorological forcings, irrigation intensity, and the aquifer. In those areas, the integration of the spatial information into MuMoC leads to an improvement of the model accuracy (NSE increased by 0.12), with peaks higher than 0.3 when the optimization objectives for selecting the neighbors were maximized.tT.

@en

The Eastern Nile Basin is facing a number of transboundary issues, including water resources development, and the associated impacts. The Nile Basin, particularly the Eastern Nile Sub-basin, is considered as one of a few international river systems of potential conflicts between riparian countries. The Eastern Nile is characterized by the high dependency of downstream countries on river water generated in upstream countries, with limited or no contribution to the runoff itself. The aim of this paper is to analyze optimal scenarios for water resources management in the Eastern Nile with regard to hydropower generation and irrigation development. A hydro-economic optimization model based on Genetic Algorithm has been used to determine the maximum benefits for two scenarios: (i) non-cooperative management of hydraulic infrastructure by the riparian countries (status quo), and (ii) cooperative water resources management among the riparian countries. The hydro-economic model is developed using a Genetic Algorithm and deterministic optimization approach covering all hydraulic infrastructures in the Eastern Nile, existing and planned, including the Grand Ethiopian Renaissance Dam (GERD). The results show that cooperative management yields an increase in hydro-energy returns for all countries compared to the status quo, with a very high increase in Ethiopian's returns, as expected. Non-cooperative system management would negatively impact the hydro-energy of Egypt compared to the cooperative management (reduced by 11%), without a significant increase of hydro-energy for Ethiopia. For Sudan, the results show that hydropower generation benefits from the presence of GERD, in both management scenarios. Non-cooperative management of the system, along with the internal trade-off between irrigation and hydropower facilities, would negatively impact irrigation supply in Sudan. The findings support the argument of positive impact of GERD development on the three Eastern Nile riparian countries, Ethiopia, Sudan and Egypt, provided that the three countries agree to manage the system cooperatively.

@en

The Eastern Nile riparian countries are currently developing several reservoir projects to contribute to the needs for energy and food production in the region. In the absence of formal mechanisms for collaboration, the transboundary nature of the Eastern Nile basin makes water resources development challenging. The large seasonal and interannual variability of the river flow increases those challenges. This paper assesses the implications of water resources development in the Eastern Nile basin on water availability for hydropower generation and irrigation demands at country and regional levels, using simulation and scenario analysis methods. Twelve scenarios are used to test developments of several dams and irrigation demands, Grand Ethiopian Renaissance Dam (GERD) operation options, and unilateral (status quo) versus integrated transboundary management of dams. A RIBASIM model that included 20 dams and 21 irrigation schemes was built, using a complete data set of 103 years at a monthly time step. Four indicators have been used for evaluating the performance of the system: hydroenergy generation (MWh=year), reliability of irrigation supply (%), reservoir net evaporation (10⁶ m³/year), and flow regimes of rivers (m³/s). The results show that in case of managing the system in an integrated transboundary manner and without new irrigation development projects, GERD would increase the hydroenergy generation in Ethiopia by + 1,500% and in Sudan by + 17%, with a slight reduction in Egypt of -1%. Supply reliability of existing and planned irrigation schemes in Sudan would not be practically influenced by the GERD, but the reliability will be reduced by about 8% when upstream development and new irrigation expansion materialize. Full development of the Eastern Nile basin would reduce the irrigation supply reliability in Egypt to 92% compared to the base scenario (100%). Compared to integrated management, unilateral management would increase the hydroenergy generation in Ethiopia (+16%), increase the rate of evaporation losses in the basin (+15%), and reduce the irrigation supply reliability in Sudan after full development of dams and irrigation projects (-10%). Water resources development would have considerable but varying effects on the countries.

@en

Intensive tillage (IT) in potato crops is considered as one of the main non-point sources (NPS) of local water eutrophication in the Fuquene Lake of Colombia. Therefore, the local government has invested in several programs aiming at the adoption of principles of conservation tillage (CT) which would allow for developing and applying the agricultural best management practices (BMPs). The complexity of hydrological and geological heterogeneity makes the degree of benefit that CT has in different locations uncertain. In this study, the Soil and Water Assessment Tool (SWAT) was used to assess the impacts of changing IT for CT on nitrogen (N) and phosphorus (P) losses in surface water runoff from the potato crop in the Fuquene watershed. This is done at field and watershed levels. A two-year study quantified the changes in surface water runoff pollutants for three potato crop cycles under the traditional IT practice and CT practice - which included reducing tillage, green manure, and permanent soil cover - at twelve runoff plots installed in the Fuquene watershed (Quintero and Comerford, 2013). This information was used to build, calibrate and validate the SWAT model. The results suggest that CT for the Fuquene watershed can be reduced up to 26% of the sediment yield and 11% of the surface runoff compared with IT, which means an overall reduction of load. The main CT effect on nutrient losses in runoff is an increase in the total N and P (2% to 18% respectively) compared to IT. However, the results at watershed level showed different patterns from those obtained at field level. Despite the model uncertainties, the results show a possibility of using hydrological models to assess the effectiveness of various field management practices in agriculture.

@en

In agricultural areas where groundwater is the main water supply for irrigation, forecasts of the water table are key to optimal water management. However, water management can be constrained by semiseasonal to seasonal forecast. The objective is to create an ensemble of water table one- to fivemonth lead-time forecasts based on multiple data-driven models (DDMs). We hypothesize that datadriven modeling capabilities (e.g., random forests, support vector machines, artificial neural networks (ANNs), extreme learning machines, and genetic programming) will improve naïve and autoregressive simulations of groundwater tables. An input variable selection method was used to carry the maximum information in the input (precipitation, crop water demand, changes in groundwater table, snowmelt, and evapotranspiration) and output relationship for the DDMs. Five DDMs were implemented to forecast groundwater tables in an unconfined aquifer in the Northern High Plains (Nebraska, USA). Root mean squared error (RMSE) and Nash-Sutcliffe efficiency index were used to evaluate the skill of the model and three hydrologic regimes were determined statistically as high, mid, and low groundwater table levels. Additionally, varying storage regimes were used to construct rising and falling limbs in the tested well. Results show that all models outperform the baseline for all the lead times, ANNs being the best of all.

@en

Increasing water demand coupled with limited water resources has given rise to the need for sustainable water resources development in the Nile River Basin. Managing the basin’s water resources is complex because of its transboundary nature, exacerbated by environmental degradation and strong climate variability. The Nile basin water resources have been extensively studied during the last 125 years for planning and management purposes, in particular with regard to the use of blue water in the downstream part of the basin, though recently some studies have also focused on the upper parts. These studies show that there is no convergence of development plans emerging among the Nile riparian countries. This paper reviews river basin water resource models as applied in the Nile River Basin, distinguishing between simulation, optimization and combined simulation and optimization models. The paper concludes by identifying knowledge gaps to guide future research on water resources planning and management in the Nile.

@en

Many terrestrial biogeochemistry process models have been applied around the world at different scales and for a large range of ecosystems. Despite being essential ecosystems that sustain important ecological processes, only a few efforts have been made to estimate the gross primary production (GPP) and the hydrological budgets along an altitudinal gradient for grasslands in the Andean Region. One of the few previous studies in the region considered the heterogeneity of the main properties of the páramo vegetation and showed significant differences in plant functional types, site/soil parameters, and daily meteorology. This study extends the work previously mentioned by using the Biome-BGC model to simulate the GPP and the water fluxes in a representative area of the Ecuadorian Andean páramos. It focuses on three main growth forms of vegetation and is also extended to cells with similar properties. The responses of GPP and the water fluxes were dependent on environmental drivers, ecophysiology, and site-specific parameters. The results showed that the GPP estimations at lower elevations are more than twice the estimations at higher elevations, which might have a large implication during extrapolations at larger spatiotemporal scales. The assessment of the water fluxes in the páramo ecosystem was inaccurate, presumably due to the poor estimation of the soil processes, water storage, and evaporative processes. A further development in the soil and evaporative modeling process of Biome-BGC is needed in order to be fully applicable in the high-altitudinal páramo ecosystems. An accurate estimation of the temporal changes of carbon and water budgets can potentially assess the effect of the climate drivers in the biomass productivity of this terrestrial ecosystem.@en

Aims The importance of quantifying carbon stocks in terrestrial ecosystems is crucial for determining climate change dynamics. However, the present regional assessments of carbon stocks in tropical grasslands are extrapolated to unsampled areas with a high degree of uncertainty and without considering the carbon and nitrogen composition of vegetation and soil along altitudinal ranges. This study aims to assess carbon and nitrogen concentrations in soil and vegetation, aboveground carbon stocks distribution and soil organic carbon stocks along an altitudinal range in the páramo region in the Ecuadorian Andes. Methods The vegetation inventory was conducted using 15×15 m sampling plots distributed in three altitudinal ranges. Based on the patterns exhibited by the dominant vegetation growth forms, biomass and soil were sampled to quantify the corresponding carbon and nitrogen concentrations. Subsequently, the aboveground live biomass along the páramo altitudinal range was estimated using allometric equations. Finally, soil and vegetation carbon stocks were estimated for the entire basin. Important Findings Altitudinal analysis supported a potential distribution of carbon and nitrogen concentrations in soil, litter and live tissues, where higher concentrations were found in the low altitudinal range mainly for tussocks and acaulescent rosettes. Cellulose in litter showed higher concentrations at low altitudinal ranges for acaulescent rosettes and cushions only. For the same growth forms, lignin patterns in litter were higher in high altitudinal ranges. Soil texture provided complementary information: high percentage of silt was highly correlated to high soil nitrogen and carbon concentration. Tussocks were found to be responsive to altitude with their, highest aboveground carbon stocks occurring at the low altitudinal range, but cushions and acaulescent rosettes responded differently. The established relationships among soil, vegetation and altitude shown in this study must be taken into account to estimate both aboveground and soil organic carbon stocks in páramo regions - such estimates will be considerably inaccurate if these relationships are ignored.

@en

As one of the main areas of carbon cycle and climate change studies, water and CO2 relations are of great significance for estimation of gross primary production (GPP). Various biogeochemical process-based models have been set up to estimate the GPP based on mathematical representation of biological, physiological and ecological processes. However, they ended up increasing the complexity and computational processing power due to the large number of physical equations that need to be solved. Computational time becomes an important matter in the simulation of multiple scenarios using models for long periods of time (e.g. climate projections). Data driven surrogate models have proven to be a useful tool for environmental modelling especially when ecological and climatic co-variates are large. The advantages of Data Driven Models (DDM) are: the possibility of adding new independent variables even if their understanding is weak, and short computational time to run. The aim is to explore the ability of DDMs to replicate a biochemical model calculating GPP. This study evaluates the performance of four surrogate DDMs, namely linear regression method (LRM), model tree (MT), instance-based learning (IBL) and artificial neural network (ANN). A simple empirical and semi-empirical relationship between GPP and climatic variables are studied. Input variable selection (IVS) methods were used to decide on the most relevant and potential environmental model inputs and then followed by a two-step approach which included a model-free and a model-based technique. Data from the highlands (páramo ecosystem) in the Ecuadorian Andean Region from 12-year time-series (2000-2011) were used to evaluate the models at various time frames and at different altitudes. The GPP time series data for the same period were derived from an earlier study using the biomodel BIOME-BGC (BioGeochemical Cycles), which is a comprehensive physical based model used in different analysis of carbon fluxes around the world. So-called IBL (nearest neighbour method) showed a great capability to reproduce the GPP when data was aggregated to monthly time frame. The computational time used to evaluate the time series with IBL as the selected DDM is shorter with enough accuracy for using it in multi-model runs.@en