Project Manager: Aliya Nugumanova, PhD in Information Systems, Director of the Big Data and Blockchain Technologies Research and Innovation Center at Astana IT University. Graduate of the Sarsen Amanzholov East Kazakhstan University. She defended her dissertation on big data and NLP at the K. Satpayev Kazakh National Research University in 2014.
Funding: THE PAF KS MSHE OF RK
Years of implementation: 2024–2026
Amount of funding: 639.4 million tenge
creation of a scientifically based system for monitoring and controlling the dynamics of surface water distribution in order to create reserve reservoirs for accumulation, regulation, and drainage of surface discharge and drainage water from temporarily flooded areas based on remote sensing data and GIS technologies.
for the first time, a set of methods will be proposed for the East Kazakhstan region to determine the estimated characteristics of the annual, maximum, and minimum runoff of unexplored rivers to solve problems of water supply, flood protection, and construction of river hydraulic structures in conditions of insufficient water resources. The established patterns of formation of water balance elements and their forecast will create a scientific basis for developing schemes for the rational use, protection, and management of water resources of the Republic of Kazakhstan.
At the global level, climate change and an increase in the frequency of extreme weather events make traditional forecasting methods less reliable. Modern methods based on remote sensing data, GIS and machine learning allow for more accurate flood forecasting, which is vital for timely response and damage minimization. The relevance of developing artificial intelligence systems that take into account complex natural and anthropogenic factors is increasing. This contributes to the development of technologies and the accumulation of new knowledge in the field of hydrology, meteorology and IT. The program stimulates interdisciplinary research that brings together scientists and specialists from various fields to solve complex flood forecasting problems.
Based on the results of the program, the following will be published:
1) at least 9 (nine) articles and (or) reviews in peer-reviewed scientific publications in the scientific area of the program, included in the 1st (first), 2nd (second) and (or) 3rd (third) quartile by impact factor in the Web of Science database and (or) having a CiteScore percentile in the Scopus database of at least 50 (fifty).
2) at least 10 (ten) articles in journals recommended by the Committee on the Study of the Scientific and Technical Education of the Russian Federation.
3) at least 1 (one) monograph or textbook in foreign and (or) Kazakhstani publishing houses recommended by the Academic Council and (or) Scientific and Technical Council of the applicant’s organization;
4) at least 2 (two) patents in foreign patent offices (European, American, Japanese) or at least 2 (two) foreign or international patents included in the Derwent Innovations Index database (Web of Science, Clarivate Analytics) or at least 5 (five) intellectual property objects (patent; for applications in the field of information technology – author’s certificate), registered with the National Institute of Intellectual Property of the Republic of Kazakhstan.
Calculation of possible accumulation reserves of melt and flood water in the territory of the East Kazakhstan region
1. Publications
Scopus:
1. Hybrid feature-based neural network regression method for load profiles forecasting
This study addresses the urgent need for improved demand forecasting models that can accurately predict energy consumption, especially under different geographical and climatic conditions. A novel demand forecasting model is proposed that combines clustering and feature engineering techniques with neural network regression, with a special focus on accounting for correlation with air temperature. The performance of the model is evaluated using a benchmark dataset from Tetouan, Morocco, where existing forecasting methods exhibit RMSE values ranging from 6,429 to 10,220 [MWh]. In contrast, the proposed model achieves a significantly lower RMSE value of 5,168, demonstrating its superiority. A subsequent application of the model to demand forecasting in Astana, Kazakhstan, as a case study further validates its performance. A comparative analysis with the baseline neural network showed a significant improvement: the proposed model achieved a MAPE value of 5.19%, while the baseline model achieved 17.36%. These results highlight the potential of the proposed approach to improve the accuracy of demand forecasting, especially in diverse geographical contexts, by using climate factors. The methodology also shows promise for broader applications such as flood forecasting, crop yield estimation, and water management.
Satan A. et al. Hybrid feature-based neural network regression method for load profiles forecasting //Energy Informatics. – 2025. – Т. 8. – №. 1. – С. 19. https://doi.org/10.1186/s42162-025-00481-0
KOKNVO:
1. Flood Risk Mapping in The Irtysh River Basin Using Satellite Data
Floods are among the most frequent and devastating natural disasters, causing significant economic losses and loss of life worldwide. Effective flood risk management depends on accurate modeling methods that can predict vulnerable areas and assess potential impacts. This study simulates flood dynamics in the Irtysh River basin near the city of Ust-Kamenogorsk (Eastern Kazakhstan), which is prone to seasonal flooding, using high-resolution satellite imagery and digital elevation models. The main objective of the work is to visually model flood risks based on terrain characteristics. The study uses satellite imagery provided by the Map box platform, which combines MODIS, Landsat 7, Maxar and Google Earth Engine data, providing access to Sentinel-2 imagery with surface reflectivity at 10-meter resolution. Elevation data from the Copernicus global digital elevation model with a resolution of 30 meters is used for flood modeling. Flood modelling involves calculating the flood depth relative to the terrain height, allowing each pixel to determine whether it will be submerged. The modelling scenarios assume a stepwise increase in water level to generate a sequence of images that show the dynamics of flooding over time. The study also considers soil hydraulic properties and focuses on visualizing flood risks based on terrain and water level changes. The modelling results show that the riverbanks are primarily affected by flooding, with water flows spreading from the north-west of the city. Critical infrastructure becomes vulnerable when the water level exceeds 2 meters from the lowest point of the terrain. These results highlight the potential of using high-resolution satellite imagery and terrain data to assess flood risks and improve urban flood preparedness. The data obtained provided valuable insights into flood development, facilitating more informed decision-making to reduce the impact of natural disasters.
Rakhymbek K., Zhomartkan N., Nurekenov D., Zhantassova Z. Flood Risk Mapping in The Irtysh River Basin Using Satellite Data //Scientific Journal of Astana IT University. – 2024. – Т. 19. – С. 140-149. https://doi.org/10.37943/19LRYW4856.
2. High-Resolution Satellite Estimation of Snow Cover for Flood Analysis in East Kazakhstan Region
The increasing frequency of extreme weather events associated with climate change makes flood forecasting particularly relevant, particularly for mountainous regions where snowmelt is the main driver of seasonal flooding. This study examines the application of snow cover assessment methods to analyze snowmelt dynamics and its potential impact on flood risks in the Ulba and Uba River basins in East Kazakhstan. To achieve this goal, high-resolution multispectral satellite imagery from the Sentinel-2 Surface Reflectance dataset is used. The analysis covers images collected from March to October for 2021–2024. Data processing is performed in the Google Earth Engine platform using strict filtering based on spatial overlap with the studied basins and the proportion of cloud pixels, which ensures high-quality data for snow cover analysis. The study applies several remote sensing indices to estimate snow cover. The Normalized Differential Snow Index (NDSI) is calculated using the green and shortwave infrared bands to identify pixels covered by snow. The fractional snow-covered area (FSCA) is calculated from the NDSI using the empirical equation ‘FRA6T’, which provides a more detailed representation of the distribution of snow across catchments. In addition, a threshold value for the ratio of near-infrared to shortwave infrared is applied, which minimizes the confusion between snow and water, especially near water bodies and during periods of active melt. The resulting snow cover maps and FSCA values provide a detailed representation of snow distribution and melt dynamics, facilitating the assessment of the role of snow runoff in flood risk development. The findings can be used to refine flood forecasting models, improve early warning systems, and support informed water management in vulnerable regions.
Alzhanov A., Nugumanova A. High-Resolution Satellite Estimation of Snow Cover for Flood Analysis in East Kazakhstan Region //Scientific Journal of Astana IT University. – 2024. – Т. 19. – С. 118-127. https://doi.org/10.37943/19VUAO6399.
3. Interrelationships between snowpack dynamics and tree growth in the Tigiretsky Ridge (Altai): Implications for ecological responses to climate variability
This study examines the complex interaction between snow cover dynamics and tree growth in the Tigiretsky Range using dendrochronological and snow cover data from 2013 to 2020. Using the Temperature-Based Melt-Index Method, the maximum snow water equivalent during winter months was accurately estimated, revealing significant spatial variability due to elevation, slope aspect, and proximity to watersheds. The results indicate an asymmetric distribution of snow cover, with southern slopes at low elevations having higher snow reserves, while this situation is reversed at higher elevations. Notably, snow reserves on northern slopes near watersheds can exceed those on southern slopes by up to 30 times. The analysis also revealed a positive correlation between the increase in snow water equivalent and the radial growth of Abies sibirica L. (Siberian fir) in the treeline ecotone, indicating a significant ecological response of trees to changing snow conditions. The findings contribute to a deeper understanding of the impact of climate variability on snow-vegetation interactions in mountain ecosystems, forming the basis for further research aimed at uncovering the mechanisms of these relationships.
Bykov N. I., Birjukov R. J. Interrelationships between snowpack dynamics and tree growth in the Tigiretsky Ridge (Altai): Implications for ecological responses to climate variability // Acta Biologica Sibirica. – 2024. –V. 10. – P. 1319–1336. https://doi.org/10.5281/zenodo.14190443 (The journal is indexed in Scopus in the direction Environmental Science -> Ecology; Global and Planetary Change, journal percentile – 29).
4. Monitoring system and provision of flood forecast data in the East Kazakhstan region
This article is devoted to the collection of available information on water bodies in the East Kazakhstan region, obtained from open sources – data provided by the Republican Hydrometeorological Service and the Ministry of Emergency Situations of the Republic of Kazakhstan. The purpose of the work is to create a spatially referenced attribute database within the GIS platform, as well as to form the basis for further research aimed at developing and implementing flood forecasting models. A review of existing studies in the field of flood forecasting was conducted both in the regions of Kazakhstan and abroad, which made it possible to formulate criteria for assessing data and hydrological monitoring systems. The article also provides a description of the hydrological regime of key representative water bodies in East Kazakhstan. Based on the spatial analysis of the network of hydrological stations and known flood zones, an assessment of the spatial coverage of the state observation system was carried out, based on the results of which recommendations were developed for its expansion. In conclusion, conclusions were made on the applicability of the collected data for building predictive models.
Pavlenko A.V., Mansurova A.K., Kyzyrkanov A., Chernykh D.V. Monitoring system and provision of flood forecast data in the East Kazakhstan region // Bulletin of the Karaganda University. Series Biology. Medicine. Geography. – 2024. – T. 4. https://doi.org/10.31489/2024bmg4/183-196
5. Feature selection methods for lstm-based river water level and discharge forecasting
Accurate forecasting of river flow and water levels is critical for effective water resources management, flood mitigation, and public safety. This study compares correlation analysis and PCA-based feature selection methods for LSTM-based forecasting models in the Uba River basin, Shemonaikha city, East Kazakhstan region. The original dataset covers the period from 1995 to 2021, with 1995–2019 data used for training and validation of the models and 2020–2021 data for testing. Both feature selection methods reduced the original set of predictors to 13 variables while maintaining overall forecast accuracy. To improve the stability of predictions and reduce variance associated with random initialization, an ensemble model of 10 LSTM networks was trained using 60-day-long input sequences and forecasting for a 10-day horizon. Model performance was assessed using the Nash–Sutcliffe Efficiency (NSE) metric. The results showed that the correlation-based feature selection method provided comparable accuracy to the model using the full feature set when tested on 2020 data, suggesting that excluding highly correlated features does not reduce the model’s short-term forecasting ability. The model with PCA-based features demonstrated some lag at longer horizons in 2020 but showed an advantage at most horizons in 2021. However, the overall forecast accuracy in 2021 decreased compared to 2020, reflecting greater variability in hydrological conditions and their deviation from historical training data, indicating the need for periodic model updates with new data. Both feature selection methods effectively reduced the data dimensionality while maintaining the predictive ability of the models. However, neither method was universally superior across all forecast time horizons. These results highlight the importance of a systematic approach to feature selection in hydrological models and the need to adapt models to changing environmental conditions.
Alzhanov A., Nugumanova A. FEATURE SELECTION METHODS FOR LSTM-BASED RIVER WATER LEVEL AND DISCHARGE FORECASTING //Scientific Journal of Astana IT University. – 2025. – Т. 21. https://doi.org/10.37943/21EHLH9882
2. Direct results
– A set of long-term data on precipitation, snow reserves, temperature conditions and solar radiation, water bodies and a set of satellite images of catchment areas and flood zones in the East Kazakhstan region have been collected.
– – Digital elevation models of catchment areas and flood zones and a conceptual model of a geo-relational database of water resources in the East Kazakhstan region have been developed.
– – A conceptual model of a geo-relational database of water resources in the East Kazakhstan region has been developed based on the ODM (Observation Data Model). PostgreSQL DBMS with the PostGIS extension, which provides support for geodata and geospatial analysis, was used for implementation.
– – 2 hydroelectric stations and 2 meteorological stations have been installed in the East Kazakhstan region.
Place of program implementation:
1. Astana – Astana IT University.
2. East Kazakhstan region, Ust-Kamenogorsk – East Kazakhstan University named after Sarsen Amanzholov.
Full name, education, degree, academic title |
H-index, ResearcherID, ORCID, Scopus Author ID (if available) |
Role in the program |
|
Nugumanova Aliya Bagdatovna, PhD in Information Systems |
H-index: Scopus: 6 |
Scientific Director, Chief Researcher |
|
Mukanova Balgaysha Gafurovna, Doctor of Physical and Mathematical Sciences |
H-index: Scopus: 6, WoS: 5 |
Chief Researcher |
|
Chernykh Dmitry Vladimirovich, Altai State University (1994), Doctor of Geographical Sciences, Associate Professor |
H-index: РИНЦ: 17, WoS: 3, Scopus: 6 |
Leading Researcher |
|
Gartsman Boris |
H-index: Scopus: 11 |
Leading Researcher |
|
Zhanasova Zheniskul |
H-index: Scopus: 2 |
Leading Researcher |
|
Moreido Vsevolod Mikhailovich, hydrologist, Candidate of Geographical Sciences |
|
Leading Researcher |
|
Bondarovich Andrey Aleksandrovich, Altai State University (1993), Candidate of Geographical Sciences, Associate Professor |
H-index: РИНЦ: 7, WoS: 3, Scopus: 4 |
Leading Researcher |
|
Bykov Nikolay Ivanovich, TSU (1984), Candidate of Geographical Sciences, Associate Professor |
H-index: РИНЦ: 15, WoS: 3, Scopus: 4 |
Senior Researcher |
|
Rakhimzhanova Anar Zhanatovna, PhD |
|
Senior Researcher |
|
Bayburin Erzhan Mukhametkalievich |
|
Senior Researcher |
|
Biryukov Roman |
H-index: Scopus: 6 |
Senior Researcher |
|
Zhakiyev Nurhat |
|
Researcher |
|
Kyzyrkanov Abzal Ermekbayuly, |
H-index: Scopus: 3 |
Researcher |
|
Ocheredko Igor |
H-index: Scopus: 4 |
Researcher |
|
Makhambetova Zhansaya Kaiyrbaevna |
ORCID: 0000-0001-5024-0289 |
Researcher |
|
|
H-index: Scopus: 1 ORCID: 0009-0007-8083-2366 |
Researcher |
|
Pavlenko Anatoly Vladimirovich, Master of Natural Sciences in Geography |
ORCID: 0000-0001-8556-6633 |
Researcher |
|
Maulit Almasbek, Master of Technical Sciences |
H-index: Scopus: 3 |
Researcher |
|
Nurekenov Dauren Makhsutbekovich, Master of Technical Sciences |
H-index: Scopus: 1 |
Researcher |
|
Mansurova Aiganym |
H-index: Scopus: 1 Scopus ID: 59233698800 ORCID 0009-0007-9076-0722 |
Researcher |
|
Zhomartkan Nurasyl Kairatuly, postgraduate student |
ORCID: 0009-0006-3935-2013 |
Researcher |
|
Rakhymbek Kamilla |
ORCID: 0009-0008-7404-8433 |
Researcher |