• Energy Research

Water scarcity poses a significant life constraint in global drylands that determines species adaptations and mosaic of exposed bare areas and vegetation patches. Runoff-water redistribution resulting from this spatial configuration has been suggested as a key process controlling water availability for vegetation and ecosystem functioning. However, the potential of this process to ameliorate the negative impacts of aridification in drylands remains unclear and there is no empirical evidence of its relevance on natural ecosystems under different levels of aridity and disturbance regimes. To address this gap, we analysed temporal series of the Normalized vegetation index (NDVI, a proxy of vegetation functioning) along a regional aridity-disturbance gradient under current and future climatic conditions. We found that mean NDVI increases in areas of runoff water accumulation (calculated using a water redistribution index) until a certain threshold, above which vegetation patches are not able to retain extra runoff water. Once threshold values were identified, we analysed the role of water redistribution on vegetation dynamics by analysing temporal series of monthly NDVI in a space for time substitution approach. The obtained results provided further evidence of the runoff water redistribution on vegetation, triggering a positive feedback between water accumulation and vegetation growth. Results obtained by the combination of the obtained model with climatic data from the 6th IPCC report, suggest that this feedback could ameliorate the expected negative effects of aridification in drylands. However, this effect is partially counterbalance in scenarios of high human disturbance and in areas where vegetation is not able to trap and retain the extra amount of resources given by runoff. Overall, our results provide empirical evidence of the relevance of runoff redistribution as a key process linking vegetation patterns to climate resistance in drylands that underscore its importance in the analysis and modelling of dryland´s responses to aridification. Funding provided by: Ministerio de Ciencia, Innovación y UniversidadesROR ID: https://ror.org/05r0vyz12Award Number: TED2021-132332BC21 Funding provided by: Ministerio de Ciencia, Innovación y UniversidadesROR ID: https://ror.org/05r0vyz12Award Number: CNS2023-144749 Funding provided by: Junta de AndalucíaROR ID: https://ror.org/01jem9c82Award Number: P18-RT-5130 Funding provided by: Ministerio de Ciencia, Innovación y UniversidadesROR ID: https://ror.org/05r0vyz12Award Number: PID2021-127631NA-I00 Funding provided by: Junta de AndalucíaROR ID: https://ror.org/01jem9c82Award Number: P_FORT_GRUPOS_2023/20 PPIT-UAL Funding provided by: Fundação para a Ciência e TecnologiaROR ID: https://ror.org/00snfqn58Award Number: DivProtect/0001/2021 Funding provided by: Agencia Estatal de InvestigaciónROR ID: https://ror.org/003x0zc53Award Number: RYC2020-030762-I

Water scarcity poses a significant life constraint in global drylands that determines species adaptations and mosaic of exposed bare areas and vegetation patches. Runoff-water redistribution resulting from this spatial configuration has been suggested as a key process controlling water availability for vegetation and ecosystem functioning. However, the potential of this process to ameliorate the negative impacts of aridification in drylands remains unclear and there is no empirical evidence of its relevance on natural ecosystems under different levels of aridity and disturbance regimes. To address this gap, we analysed temporal series of the Normalized vegetation index (NDVI, a proxy of vegetation functioning) along a regional aridity-disturbance gradient under current and future climatic conditions. We found that mean NDVI increases in areas of runoff water accumulation (calculated using a water redistribution index) until a certain threshold, above which vegetation patches are not able to retain extra runoff water. Once threshold values were identified, we analysed the role of water redistribution on vegetation dynamics by analysing temporal series of monthly NDVI in a space for time substitution approach. The obtained results provided further evidence of the runoff water redistribution on vegetation, triggering a positive feedback between water accumulation and vegetation growth. Results obtained by the combination of the obtained model with climatic data from the 6th IPCC report, suggest that this feedback could ameliorate the expected negative effects of aridification in drylands. However, this effect is partially counterbalance in scenarios of high human disturbance and in areas where vegetation is not able to trap and retain the extra amount of resources given by runoff. Overall, our results provide empirical evidence of the relevance of runoff redistribution as a key process linking vegetation patterns to climate resistance in drylands that underscore its importance in the analysis and modelling of dryland´s responses to aridification. Funding provided by: Ministerio de Ciencia, Innovación y UniversidadesROR ID: https://ror.org/05r0vyz12Award Number: TED2021-132332BC21 Funding provided by: Ministerio de Ciencia, Innovación y UniversidadesROR ID: https://ror.org/05r0vyz12Award Number: CNS2023-144749 Funding provided by: Junta de AndalucíaROR ID: https://ror.org/01jem9c82Award Number: P18-RT-5130 Funding provided by: Ministerio de Ciencia, Innovación y UniversidadesROR ID: https://ror.org/05r0vyz12Award Number: PID2021-127631NA-I00 Funding provided by: Junta de AndalucíaROR ID: https://ror.org/01jem9c82Award Number: P_FORT_GRUPOS_2023/20 PPIT-UAL Funding provided by: Fundação para a Ciência e TecnologiaROR ID: https://ror.org/00snfqn58Award Number: DivProtect/0001/2021 Funding provided by: Agencia Estatal de InvestigaciónROR ID: https://ror.org/003x0zc53Award Number: RYC2020-030762-I

Water availability controls the functioning of dryland ecosystems, driving a patchy vegetation distribution, unequal nutrient availability, soil respiration in pulses, and limited productivity. Groundwater-dependent ecosystems (GDEs) are acknowledged to be decoupled from precipitation, since their vegetation relies on groundwater sources. Despite their relevance to enhance productivity in drylands, our understanding of how different components of GDEs interconnect (i.e., soil, vegetation, water) remains limited. We studied the GDE dominated by the deep-rooted phreatophyte Ziziphus lotus, a winter-deciduous shrub adapted to arid conditions along the Mediterranean basin. We aimed to disentangle whether the groundwater connection established by Z. lotus will foster soil biological activity and therefore soil fertility in drylands. We assessed (1) soil and vegetation dynamics over seasons (soil CO2 efflux and plant activity), (2) the effect of the patchy distribution on soil quality (properties and nutrient availability), and soil biological activity (microbial biomass and mineralization rates) as essential elements of biogeochemical cycles, and (3) the implications for preserving GDEs and their biogeochemical processes under climate change effects. We found that soil and vegetation dynamics respond to water availability. Whereas soil biological activity promptly responded to precipitation events, vegetation functioning relies on less superficial water and responded on different time scales. Soil quality was higher under the vegetation patches, as was soil biological activity. Our findings highlight the importance of groundwater connections and phreatophytic vegetation to increase litter inputs and organic matter into the soils, which in turn enhances soil quality and decomposition processes in drylands. However, biogeochemical processes are jeopardized in GDEs by climate change effects and land degradation due to the dependence of soil activity on: (1) precipitation for activation, and (2) phreatophytic vegetation for substrate accumulation. Therefore, desertification might modify biogeochemical cycles by disrupting key ecosystem processes such as soil microbial activity, organic matter mineralization, and plant productivity.

Water availability controls the functioning of dryland ecosystems, driving a patchy vegetation distribution, unequal nutrient availability, soil respiration in pulses, and limited productivity. Groundwater-dependent ecosystems (GDEs) are acknowledged to be decoupled from precipitation, since their vegetation relies on groundwater sources. Despite their relevance to enhance productivity in drylands, our understanding of how different components of GDEs interconnect (i.e., soil, vegetation, water) remains limited. We studied the GDE dominated by the deep-rooted phreatophyte Ziziphus lotus, a winter-deciduous shrub adapted to arid conditions along the Mediterranean basin. We aimed to disentangle whether the groundwater connection established by Z. lotus will foster soil biological activity and therefore soil fertility in drylands. We assessed (1) soil and vegetation dynamics over seasons (soil CO2 efflux and plant activity), (2) the effect of the patchy distribution on soil quality (properties and nutrient availability), and soil biological activity (microbial biomass and mineralization rates) as essential elements of biogeochemical cycles, and (3) the implications for preserving GDEs and their biogeochemical processes under climate change effects. We found that soil and vegetation dynamics respond to water availability. Whereas soil biological activity promptly responded to precipitation events, vegetation functioning relies on less superficial water and responded on different time scales. Soil quality was higher under the vegetation patches, as was soil biological activity. Our findings highlight the importance of groundwater connections and phreatophytic vegetation to increase litter inputs and organic matter into the soils, which in turn enhances soil quality and decomposition processes in drylands. However, biogeochemical processes are jeopardized in GDEs by climate change effects and land degradation due to the dependence of soil activity on: (1) precipitation for activation, and (2) phreatophytic vegetation for substrate accumulation. Therefore, desertification might modify biogeochemical cycles by disrupting key ecosystem processes such as soil microbial activity, organic matter mineralization, and plant productivity.

# Resource redistribution mediated by hydrological connectivity modulates vegetation response to aridification in drylands [https://doi.org/10.5061/dryad.kwh70rzfk](https://doi.org/10.5061/dryad.kwh70rzfk) ## Description of the data and file structure We used R software for data analysis. Long term climatic data (average values for the period 1970-2000) were obtained from the Spanish Climate Atlas developed by the Spanish Meteorological Agency (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)). Daily climatic data for the studied period was obtainded from the climatic network of the Instituto Valenciano de Investigaciones Agrarias (IVIA, [http://riegos.ivia.es/listado-de-estaciones](http://riegos.ivia.es/listado-de-estaciones)) and the Instituto Murciano de Investigación y Desarrollo Agrario y Medioambiental (IMIDA, [http://siam.imida.es](http://siam.imida.es)). Climate predictions for the 6th IPCC report were downloaded from [https://esgf-ui.ceda.ac.uk/cog/search/cmip6-ceda/](https://esgf-ui.ceda.ac.uk/cog/search/cmip6-ceda/). Digital Elevation Models (DEMs) were obtained from the Instituto Geográfico Nacional (IGN, [https://www.ign.es/web/ign/portal](https://www.ign.es/web/ign/portal)). Orthoimages were obtained from National Plan of Aerial Orthophotography from Spain (PNOA, [https://pnoa.ign.es/](https://pnoa.ign.es/)). Aridity values were obtained from Global Aridity Index Database (Trabucco and Zomer, 2018). Global Human Influence Index v2 was obtained from [https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic](https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic) . NDVI from Sentinel-2 imagery was calculated in Google Earth Engine code editor ([https://code.earthengine.google.com/](https://code.earthengine.google.com/)) using the COPERNICUS/S2 image collection. --- **General Glossary**: GAI: Global aridity index (Trabucco and Zomer, 2018) ETP: Potential Evapotranspiration (Trabucco and Zomer, 2018) Tmax: Mean annual maximum temperature (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) Tmean: Mean annual temperature (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) PP: Mean annual precipitation (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) COVER: Vegetation cover (Calculated from PNOA, [https://pnoa.ign.es/](https://pnoa.ign.es/)) NDVI: Normalized Difference Vegetation Index (Calcutated from Sentinel-2 images). HI: Global Human Influence Index ([https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic](https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic)) retention: retentioni from BalanR, a water redistribution index from (Rodriguez-Lozano et al., 2023) --- **Dataset description:** POINTID: Identifier of each pixel from SENTINEL-2 images studied CODE: Identifier of each study area CodYear: Identifier of each hydrological year studied NDVI: Normalized Difference Vegetation Index (Calcutated from Sentinel-2 images). Can contain NA due to cloud masking and low data quality removal. NDVIn-1: NDVI in the previous month. Please note that for the first month of data NDVIn-1 is NA. In addition, it can contain NA due to cloud masking and low data quality removal. PP_n-1 to PP_9month: Precipitation in the previous month to 9 months before NDVI adquisition. This information was obtained from Instituto Valenciano de Investigaciones Agrarias (IVIA, [http://riegos.ivia.es/listado-de-estaciones](http://riegos.ivia.es/listado-de-estaciones)) and the Instituto Murciano de Investigación y Desarrollo Agrario y Medioambiental (IMIDA, [http://siam.imida.es](http://siam.imida.es)). Solar: Potential incoming solar radiation calculated with the hemispherical visual basin algorithm with 100% of atmospheric transmissivity in ArcGis 10.1 (Fu, 2000). Altitude: Altitude (m.a.s.l.) of each study area COVER: Vegetation cover of each study pixel. Please note that you need to multiply by 100 to obtain the % Plot_Altitude: Mean altitude of each study plot Plot_COVER: Mean vegetation cover of each study plot. Please note that you need to multiply by 100 to obtain the % GAI:Global aridity index (Trabucco and Zomer, 2018) Plot_PP, Plot_Tmax and Plot_Tmean: Mean annual precipitation, Mean annual maximum temperature and Mean annual temperature (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) of each study plot Plot_HI: Mean Global Human Influence Index ([https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic](https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic)) of each plot ETP: Potential Evapotranspiration (Trabucco and Zomer, 2018) ETP_1month to ETP_9month: Potential Evapotranspiration in the previous month to 9 months before NDVI adquisition. This information was calculated based on potential solar radiation and climatic data from Instituto Valenciano de Investigaciones Agrarias (IVIA, [http://riegos.ivia.es/listado-de-estaciones](http://riegos.ivia.es/listado-de-estaciones)) and the Instituto Murciano de Investigación y Desarrollo Agrario y Medioambiental (IMIDA, [http://siam.imida.es](http://siam.imida.es)). WB1 to WB9: Water balance from month he previous month to 9 months before NDVI adquisition. This information was calculated based on ETP_1month to ETP_9month and PP_n-1 to PP_9month class: Identifies the pixels below and above the thresholds (breakpoints). Water scarcity poses a significant life constraint in global drylands that determines species adaptations and mosaic of exposed bare areas and vegetation patches. Runoff-water redistribution resulting from this spatial configuration has been suggested as a key process controlling water availability for vegetation and ecosystem functioning. However, the potential of this process to ameliorate the negative impacts of aridification in drylands remains unclear and there is no empirical evidence of its relevance on natural ecosystems under different levels of aridity and disturbance regimes. To address this gap, we analysed temporal series of the Normalized vegetation index (NDVI, a proxy of vegetation functioning) along a regional aridity-disturbance gradient under current and future climatic conditions. We found that mean NDVI increases in areas of runoff water accumulation (calculated using a water redistribution index) until a certain threshold, above which vegetation patches are not able to retain extra runoff water. Once threshold values were identified, we analysed the role of water redistribution on vegetation dynamics by analysing temporal series of monthly NDVI in a space for time substitution approach. The obtained results provided further evidence of the runoff water redistribution on vegetation, triggering a positive feedback between water accumulation and vegetation growth. Results obtained by the combination of the obtained model with climatic data from the 6th IPCC report, suggest that this feedback could ameliorate the expected negative effects of aridification in drylands. However, this effect is partially counterbalance in scenarios of high human disturbance and in areas where vegetation is not able to trap and retain the extra amount of resources given by runoff. Overall, our results provide empirical evidence of the relevance of runoff redistribution as a key process linking vegetation patterns to climate resistance in drylands that underscore its importance in the analysis and modelling of dryland´s responses to aridification.

# Resource redistribution mediated by hydrological connectivity modulates vegetation response to aridification in drylands [https://doi.org/10.5061/dryad.kwh70rzfk](https://doi.org/10.5061/dryad.kwh70rzfk) ## Description of the data and file structure We used R software for data analysis. Long term climatic data (average values for the period 1970-2000) were obtained from the Spanish Climate Atlas developed by the Spanish Meteorological Agency (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)). Daily climatic data for the studied period was obtainded from the climatic network of the Instituto Valenciano de Investigaciones Agrarias (IVIA, [http://riegos.ivia.es/listado-de-estaciones](http://riegos.ivia.es/listado-de-estaciones)) and the Instituto Murciano de Investigación y Desarrollo Agrario y Medioambiental (IMIDA, [http://siam.imida.es](http://siam.imida.es)). Climate predictions for the 6th IPCC report were downloaded from [https://esgf-ui.ceda.ac.uk/cog/search/cmip6-ceda/](https://esgf-ui.ceda.ac.uk/cog/search/cmip6-ceda/). Digital Elevation Models (DEMs) were obtained from the Instituto Geográfico Nacional (IGN, [https://www.ign.es/web/ign/portal](https://www.ign.es/web/ign/portal)). Orthoimages were obtained from National Plan of Aerial Orthophotography from Spain (PNOA, [https://pnoa.ign.es/](https://pnoa.ign.es/)). Aridity values were obtained from Global Aridity Index Database (Trabucco and Zomer, 2018). Global Human Influence Index v2 was obtained from [https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic](https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic) . NDVI from Sentinel-2 imagery was calculated in Google Earth Engine code editor ([https://code.earthengine.google.com/](https://code.earthengine.google.com/)) using the COPERNICUS/S2 image collection. --- **General Glossary**: GAI: Global aridity index (Trabucco and Zomer, 2018) ETP: Potential Evapotranspiration (Trabucco and Zomer, 2018) Tmax: Mean annual maximum temperature (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) Tmean: Mean annual temperature (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) PP: Mean annual precipitation (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) COVER: Vegetation cover (Calculated from PNOA, [https://pnoa.ign.es/](https://pnoa.ign.es/)) NDVI: Normalized Difference Vegetation Index (Calcutated from Sentinel-2 images). HI: Global Human Influence Index ([https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic](https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic)) retention: retentioni from BalanR, a water redistribution index from (Rodriguez-Lozano et al., 2023) --- **Dataset description:** POINTID: Identifier of each pixel from SENTINEL-2 images studied CODE: Identifier of each study area CodYear: Identifier of each hydrological year studied NDVI: Normalized Difference Vegetation Index (Calcutated from Sentinel-2 images). Can contain NA due to cloud masking and low data quality removal. NDVIn-1: NDVI in the previous month. Please note that for the first month of data NDVIn-1 is NA. In addition, it can contain NA due to cloud masking and low data quality removal. PP_n-1 to PP_9month: Precipitation in the previous month to 9 months before NDVI adquisition. This information was obtained from Instituto Valenciano de Investigaciones Agrarias (IVIA, [http://riegos.ivia.es/listado-de-estaciones](http://riegos.ivia.es/listado-de-estaciones)) and the Instituto Murciano de Investigación y Desarrollo Agrario y Medioambiental (IMIDA, [http://siam.imida.es](http://siam.imida.es)). Solar: Potential incoming solar radiation calculated with the hemispherical visual basin algorithm with 100% of atmospheric transmissivity in ArcGis 10.1 (Fu, 2000). Altitude: Altitude (m.a.s.l.) of each study area COVER: Vegetation cover of each study pixel. Please note that you need to multiply by 100 to obtain the % Plot_Altitude: Mean altitude of each study plot Plot_COVER: Mean vegetation cover of each study plot. Please note that you need to multiply by 100 to obtain the % GAI:Global aridity index (Trabucco and Zomer, 2018) Plot_PP, Plot_Tmax and Plot_Tmean: Mean annual precipitation, Mean annual maximum temperature and Mean annual temperature (AEMET, [http://agroclimap.aemet.es](http://agroclimap.aemet.es)) of each study plot Plot_HI: Mean Global Human Influence Index ([https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic](https://sedac.ciesin.columbia.edu/data/set/wildareas-v2-human-influence-index-geographic)) of each plot ETP: Potential Evapotranspiration (Trabucco and Zomer, 2018) ETP_1month to ETP_9month: Potential Evapotranspiration in the previous month to 9 months before NDVI adquisition. This information was calculated based on potential solar radiation and climatic data from Instituto Valenciano de Investigaciones Agrarias (IVIA, [http://riegos.ivia.es/listado-de-estaciones](http://riegos.ivia.es/listado-de-estaciones)) and the Instituto Murciano de Investigación y Desarrollo Agrario y Medioambiental (IMIDA, [http://siam.imida.es](http://siam.imida.es)). WB1 to WB9: Water balance from month he previous month to 9 months before NDVI adquisition. This information was calculated based on ETP_1month to ETP_9month and PP_n-1 to PP_9month class: Identifies the pixels below and above the thresholds (breakpoints). Water scarcity poses a significant life constraint in global drylands that determines species adaptations and mosaic of exposed bare areas and vegetation patches. Runoff-water redistribution resulting from this spatial configuration has been suggested as a key process controlling water availability for vegetation and ecosystem functioning. However, the potential of this process to ameliorate the negative impacts of aridification in drylands remains unclear and there is no empirical evidence of its relevance on natural ecosystems under different levels of aridity and disturbance regimes. To address this gap, we analysed temporal series of the Normalized vegetation index (NDVI, a proxy of vegetation functioning) along a regional aridity-disturbance gradient under current and future climatic conditions. We found that mean NDVI increases in areas of runoff water accumulation (calculated using a water redistribution index) until a certain threshold, above which vegetation patches are not able to retain extra runoff water. Once threshold values were identified, we analysed the role of water redistribution on vegetation dynamics by analysing temporal series of monthly NDVI in a space for time substitution approach. The obtained results provided further evidence of the runoff water redistribution on vegetation, triggering a positive feedback between water accumulation and vegetation growth. Results obtained by the combination of the obtained model with climatic data from the 6th IPCC report, suggest that this feedback could ameliorate the expected negative effects of aridification in drylands. However, this effect is partially counterbalance in scenarios of high human disturbance and in areas where vegetation is not able to trap and retain the extra amount of resources given by runoff. Overall, our results provide empirical evidence of the relevance of runoff redistribution as a key process linking vegetation patterns to climate resistance in drylands that underscore its importance in the analysis and modelling of dryland´s responses to aridification.