Open AccessArticle Assessment of Ecological Sensitivity to Climate Change in Southern Kazakhstan: A Composite NDVI–Climate Index Approach (2010–2025) 1 Department of Ecology, M.Auezov South Kazakhstan University, Shymkent 160012, Kazakhstan 2 Department of Water Resources, Land Using and Agrotechnology, M.Auezov South Kazakhstan University, Shymkent 160012, Kazakhstan 3 Department of Architecture and Urban planning, M.Auezov South Kazakhstan University, Shymkent 160012, Kazakhstan * Authors to whom correspondence should be addressed. Diversity 2026, 18(6), 347; https://doi.org/10.3390/d18060347 (registering DOI) Submission received: 12 May 2026 / Revised: 29 May 2026 / Accepted: 5 June 2026 / Published: 7 June 2026 Abstract Climate change threatens ecosystem stability in arid Central Asia, yet regional vegetation responses remain poorly resolved at the operational scale of land-use policy. We integrated long-term meteorological records (2000–2024) from Kazhydromet with Landsat surface-reflectance imagery for four epochs (2010, 2015, 2020, 2025) across the five administrative regions of Southern Kazakhstan (≈710,000 km 2). After cross-sensor harmonization of Landsat 5 TM and Landsat 8 OLI, dense vegetation cover (NDVI > 0.4) increased modestly across all regions, with the cumulative area growing from 9.09 to 9.60 million hectares (+5.6%) and a transient 2020 minimum linked to the 2018–2020 drought. Per-region OLS trend slopes were not statistically significant at p 0.57)—using natural-breaks classification optimized for the regional BCSI distribution. To assess robustness, the BCSI was recomputed under three alternative weighting schemes: climate-dominant (w 1 = 0.40, w 2 = 0.30, w 3 = 0.20, w 4 = 0.10); vegetation-dominant (w 1 = 0.20, w 2 = 0.10, w 3 = 0.40, w 4 = 0.30); and precipitation-dominant. Across all alternatives, the spatial location of Very High and Very Low classes was preserved ( Section 3.3). Note on terminology. The term ‘Biodiversity–Climate Sensitivity Index’ is retained for compatibility with related regional-scale composite-index frameworks; however, in the present study, the biodiversity component (X 4) is operationalized through the coefficient of variation of remotely sensed dense-vegetation cover, which is a proxy for ecosystem vulnerability rather than a direct measurement of species richness or diversity. Direct species-level integration is identified as a priority for future versions of the index ( Section 4.5). 2.5. Analytical Workflow The integrated workflow consisted of (i) data preprocessing and harmonization, including cross-sensor calibration of Landsat 5/8; (ii) temporal trend analysis of climatic and NDVI variables using parallel OLS and Mann–Kendall procedures; (iii) per-region computation of dense-vegetation cover statistics (mean, standard deviation, coefficient of variation, OLS slope, R 2, drought response, recovery); (iv) BCSI computation with equal weights and sensitivity testing under three alternative weighting schemes; and (v) comparative evaluation of ecosystem responses across the five administrative regions and the three broad ecosystem types (desert, steppe, mountain). The full analytical framework is summarized in Figure 2. 3. Results 3.1. Climatic Context, 2000–2024 3.2. Spatial NDVI Patterns and Dense-Vegetation Dynamics Spatial NDVI distributions for the four target years are shown in Figure 3 as a five-class gradient ranging from water bodies and bare ground (NDVI ≤ 0.0) to dense vegetation (NDVI > 0.6). Across all four epochs, the highest NDVI values are concentrated in the foothill and mountain belts of the Almaty and Zhetysu regions and in the irrigated belt of central Turkistan. In contrast, the desert lowlands of Kyzylorda and parts of Zhambyl maintain low NDVI ( 0.4 used to delineate dense vegetation is a conservative choice appropriate for Central Asian drylands but may underestimate productivity in sparse pastures and semi-deserts; a continuous-NDVI metric (e.g., growing-season-integrated NDVI) would provide a more nuanced productivity signal. Third, the biodiversity component of the BCSI (X 4) is currently operationalized using the coefficient of variation of dense-vegetation cover, a remote-sensing proxy for ecosystem vulnerability rather than a direct measure of species-level diversity. Direct integration of species richness (S) and Shannon diversity (H) data—for example, from the Kazakhstan National Biodiversity Information System, the Republic Red Data Book of Kazakhstan, or open platforms such as GBIF—is identified as a priority for future iterations of the index. Fourth, the analysis does not explicitly disentangle climatic drivers from anthropogenic pressures, such as grazing intensity, irrigation withdrawal, and land-use change, all of which strongly influence vegetation dynamics in Southern Kazakhstan. Recent regional assessments document substantial pollution loads from municipal solid waste in the Shymkent–Turkistan agglomeration [ 16]. Integrating such anthropogenic-pressure layers is an additional priority. Decomposing greening pixels using a global irrigated-cropland mask, such as GFSAD30 [ 31], would directly quantify the management contribution to the observed signal. Finally, the BCSI relies on equal weights as a defensible default; sensitivity analyses with three alternative weighting schemes ( Section 3.3) confirmed that the spatial pattern of Very High and Very Low classes is preserved, but quantitative BCSI values shift modestly. Formal weight derivation through expert elicitation (e.g., the Analytic Hierarchy Process) would further strengthen the index but is reserved for future versions that also integrate species-level biodiversity data. These limitations define priority directions for future research and should be considered when transferring the proposed framework to other arid and semi-arid regions. 5. Conclusions This study presents an integrated, regionally resolved assessment of climate–vegetation interactions in the ecosystems of Southern Kazakhstan from 2010 to 2025, combining open-access climate data from Kazhydromet, multi-temporal Landsat NDVI imagery harmonized across sensors, and a composite Biodiversity–Climate Sensitivity Index (BCSI). Three findings are central. First, the multi-temporal NDVI analysis reveals ecosystem-specific patterns. Mountain ecosystems in the Almaty and Zhetysu regions had the highest dense-vegetation cover (means of 30.54% and 26.27%, respectively) and the lowest inter-epoch volatility (CV ≈ 2.0%). Steppe and foothill landscapes in Zhambyl and Turkistan exhibited moderate but stable coverage. Arid ecosystems in Kyzylorda, despite the lowest absolute vegetation cover (4.66%), showed unexpected stability with no statistically significant decline. The 2020 drought response was strongest in arid lowlands (−2.0 to −2.5%) and weakest in mountains (−0.8%), and the 2020–2025 recovery followed the same gradient. Second, the cumulative dense-vegetation area in Southern Kazakhstan increased from 9.092 Mha in 2010 to 9.602 Mha in 2025—a net change of +5.6% over 15 years—with a transient minimum of 9.385 Mha in 2020. Every administrative region recorded a positive net change, ranging from +0.29 percentage points in Kyzylorda to +1.63 pp in Almaty. Importantly, none of the per-region OLS trends are statistically significant at p 0.6), including irrigated croplands and forests. All maps use the same projection (WGS 1984 UTM Zone 42N) and legend. The reduction in moderate NDVI in 2020 reflects drought conditions during 2018–2020. Figure 3. Spatial distribution of NDVI across Southern Kazakhstan for ( a) 2010, ( b) 2015, ( c) 2020, and ( d) 2025 derived from harmonized Landsat 5/8 surface reflectance composites. NDVI is classified into five ecosystem types: water/bare (≤0.0), desert/built-up (0.0–0.2), sparse vegetation (0.2–0.4), moderate vegetation (0.4–0.6), and dense vegetation (>0.6), including irrigated croplands and forests. All maps use the same projection (WGS 1984 UTM Zone 42N) and legend. The reduction in moderate NDVI in 2020 reflects drought conditions during 2018–2020. Figure 4. Spatial distribution of dense vegetation cover (NDVI > 0.4) in Southern Kazakhstan for ( a) 2010, ( b) 2015, ( c) 2020, and ( d) 2025, showing temporal changes in vegetation extent across the study period. Figure 4. Spatial distribution of dense vegetation cover (NDVI > 0.4) in Southern Kazakhstan for ( a) 2010, ( b) 2015, ( c) 2020, and ( d) 2025, showing temporal changes in vegetation extent across the study period. Figure 5. Composite Biodiversity–Climate Sensitivity Index (BCSI) in Southern Kazakhstan, 2025. Panels ( a– e) represent Kyzylorda, Turkistan, Zhambyl, Almaty, and Zhetysu regions. BCSI is grouped into five classes from Very Low ( 0.57), with higher values indicating greater ecosystem sensitivity. Insets show regional locations in Kazakhstan. Figure 5. Composite Biodiversity–Climate Sensitivity Index (BCSI) in Southern Kazakhstan, 2025. Panels ( a– e) represent Kyzylorda, Turkistan, Zhambyl, Almaty, and Zhetysu regions. BCSI is grouped into five classes from Very Low ( 0.57), with higher values indicating greater ecosystem sensitivity. Insets show regional locations in Kazakhstan. Table 1. Summary of climatic context for Southern Kazakhstan, 2000–2024. Table 1. Summary of climatic context for Southern Kazakhstan, 2000–2024. Indicator Reported Pattern, 2000–2024 Direction Source Note: Arrows indicate the direction of change (↑ increase, ↓ decrease, → stable/no significant change). Table 2. Ecosystem-level vegetation indicators across regions of Southern Kazakhstan (2010–2025), derived from Landsat Collection 2 dense-vegetation time series (NDVI > 0.4). Table 2. Ecosystem-level vegetation indicators across regions of Southern Kazakhstan (2010–2025), derived from Landsat Collection 2 dense-vegetation time series (NDVI > 0.4). Region Ecosystem Type Mean Cover 2010–2025 (%) SD (pp) CV (%) OLS Slope (pp/dec) R 22020 Drought Response (%) 2020 → 2025 Recovery (%) Almaty Mountain 30.54 0.63 2.05 0.93 0.691 −0.78 +1.57 Zhetysu Mountain/Steppe 26.27 0.52 1.97 0.76 0.675 −0.79 +1.44 Turkistan Steppe/Semi-desert 8.73 0.16 1.78 0.20 0.508 −2.05 +4.07 Zhambyl Semi-desert/Desert 6.95 0.15 2.09 0.20 0.583 −2.13 +3.48 Kyzylorda Desert 4.66 0.11 2.38 0.15 0.571 −2.54 +4.35 Note: Mean = mean share of dense-vegetation cover (NDVI > 0.4) averaged over four epochs (2010, 2015, 2020, 2025). SD = standard deviation in percentage points. CV = coefficient of variation (%). OLS slope estimated by ordinary least-squares regression ( n = 4); none of the slopes is statistically significant at p 0.57 Concentrated in the lower Syr Darya floodplain and former Aral Sea margins; pervasive desertification signal. Highest Turkistan Moderate (mixed) 0.50–0.54 Low–Moderate in the central irrigated belt; elevated sensitivity along the Kyzylkum and Moyynkum desert margins High Zhambyl Moderate 0.50–0.54 Low-sensitivity foothill belt of the Kyrgyz Range; the Talas and Chu river corridors are more resilient. Medium Almaty Bimodal: Very Low/Moderate <0.30 and 0.50–0.54 The mountainous SE (Northern Tien Shan) is the most resilient zone in the study area; the lowland and reservoir margins are sensitive. Lowest (mtn)/Medium (lowland) Zhetysu Low to Moderate 0.30–0.54 The foothills of the Dzhungarian Alatau are resilient, with a gradient toward the Balkhash drainage. Medium-low Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. © 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license. Share and Cite MDPI and ACS Style Abduova, A.; Kaldybek, E.; Kenzhaliyeva, G.; Bektureyeva, G.; Zhorabayeva, N.; Yussupova, A.; Kozhakhmetova, A.; Askerbekova, A.; Tileuberdi, A.; Sabyrkhan, A. Assessment of Ecological Sensitivity to Climate Change in Southern Kazakhstan: A Composite NDVI–Climate Index Approach (2010–2025). Diversity 2026, 18, 347. https://doi.org/10.3390/d18060347 AMA Style Abduova A, Kaldybek E, Kenzhaliyeva G, Bektureyeva G, Zhorabayeva N, Yussupova A, Kozhakhmetova A, Askerbekova A, Tileuberdi A, Sabyrkhan A. Assessment of Ecological Sensitivity to Climate Change in Southern Kazakhstan: A Composite NDVI–Climate Index Approach (2010–2025). Diversity. 2026; 18(6):347. https://doi.org/10.3390/d18060347 Chicago/Turabian Style Abduova, Aisulu, Erzhan Kaldybek, Gulmira Kenzhaliyeva, Gulzhan Bektureyeva, Nailya Zhorabayeva, Akmaral Yussupova, Aidana Kozhakhmetova, Arailym Askerbekova, Ayaulym Tileuberdi, and Arailym Sabyrkhan. 2026. "Assessment of Ecological Sensitivity to Climate Change in Southern Kazakhstan: A Composite NDVI–Climate Index Approach (2010–2025)" Diversity 18, no. 6: 347. https://doi.org/10.3390/d18060347 APA Style Abduova, A., Kaldybek, E., Kenzhaliyeva, G., Bektureyeva, G., Zhorabayeva, N., Yussupova, A., Kozhakhmetova, A., Askerbekova, A., Tileuberdi, A., & Sabyrkhan, A. (2026). Assessment of Ecological Sensitivity to Climate Change in Southern Kazakhstan: A Composite NDVI–Climate Index Approach (2010–2025). Diversity, 18(6), 347. https://doi.org/10.3390/d18060347 Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details . Article Metrics Article metric data becomes available approximately 24 hours after publication online.
