Ten rock samples consisting of one pyroclastic density current(PDC1)deposit,seven lava flows(LF1–7),and two summit lava domes(LD1,2)were studied to understand the petrogenesis and magma dynamics at Mt.Sumbing.The str...Ten rock samples consisting of one pyroclastic density current(PDC1)deposit,seven lava flows(LF1–7),and two summit lava domes(LD1,2)were studied to understand the petrogenesis and magma dynamics at Mt.Sumbing.The stratigraphy is arranged as LF1,PDC1,LF2,LF3,LF4,LF5,LF6,LF7,LD1,and LD2;furthermore,these rocks were divided into two types.TypeⅠ,observed in the oldest(LF1)sample,has poor MgO and high Ba/Nb,Th/Yb and Sr.The remaining samples(PDC1–LD2)represent typeⅡ,characterized by high MgO and low Ba/Nb,Th/Yb and Sr values.We suggest that type I is derived from AOC(altered oceanic crust)-rich melts that underwent significant crustal assimilation,while typeⅡoriginates from mantle-rich melts with less significant crustal assimilation.The early stage of typeⅡmagma(PDC1–LF3)was considered a closed system,evolving basaltic andesite into andesite(55.0–60.2 wt%SiO_(2))with a progressively increasing phenocryst(0.30–0.48φ_(PC))and decreasing crystal size distribution(CSD)slope(from-3.9 to-2.9).The evidence of fluctuating silica and phenocryst contents(between 55.9–59.7 wt%and 0.25–0.41φ_(PC),respectively),coupled with the kinked and steep(from-5.0 to-3.3)CSD curves imply the interchanging condition between open(i.e.,magma mixing)and closed magmatic systems during the middle stage(LF4–LF6).Finally,it underwent to closed system again during the final stage(LF7–LD2)because the magma reached dacitic composition(at most 68.9 wt%SiO_(2))with abundant phenocryst(0.38–0.45φ_(PC))and gentle CSD slope(from-4.1 to-1.2).展开更多
High concentration ground-level ozone(O3)has adverse effects on plant growth and photosynthesis.Compared to the O3concentration-based index,the O3flux-based(especially stomatal O3uptake)index has been considered the b...High concentration ground-level ozone(O3)has adverse effects on plant growth and photosynthesis.Compared to the O3concentration-based index,the O3flux-based(especially stomatal O3uptake)index has been considered the better criterion for assessing the impact of ozone on vegetation and ecosystems.This paper reports on a study of O3flux using the eddy covariance technique over a corn field in the Northwestern Shandong Plain of China.Diurnal variation of atmospheric O3concentration,deposition velocity and flux,and their relationships to environmental factors are analyzed.The results show that:(1)During the observation period(9 August–28 September,2011),there was a strong diurnal variation of O3concentration,with low(16.5 nL L?1)and high(60.1 nL L?1)O3mean concentrations observed around 6:30 and 16:00,respectively.Mean O3concentrations during daytime(6:00–18:00)and nighttime(18:00–6:00)were 39.8±23.1 and 20.7±14.1 nL L?1(mean±std),respectively.The maximum observed concentration was 97.5 nL L?1.The concentration was mainly affected by solar radiation and air temperature.(2)Whether daytime or nighttime,ground-level O3flux is always downward.The diurnal course of mean deposition velocity was divided into 4 phases:a low and stable process during nighttime,fast increasing in early morning,relatively large and steady changes around noon,and quickly decreasing in later afternoon.Daytime and nighttime mean deposition velocities were 0.29 and 0.09 cm s?1,respectively.The maximum deposition velocity was 0.81 cm s?1.The magnitude of deposition velocity was influenced by the corn growth period,and its diurnal variation was significantly correlated with global radiation and relative humidity.(3)O3flux was affected by variations of both O3concentration and deposition velocity,with mean O3fluxes-317.7 and?70.2 ng m?2s?1during daytime and nighttime,respectively.There was strong correlation between O3flux and CO2flux or latent heat flux.By comparing the deposition velocities of daytime and nighttime,we infer that stomatal uptake was probably the main sink of ground-level O3.展开更多
Machine learning(ML)algorithms are frequently used in landslide susceptibility modeling.Different data handling strategies may generate variations in landslide susceptibility modeling,even when using the same ML algor...Machine learning(ML)algorithms are frequently used in landslide susceptibility modeling.Different data handling strategies may generate variations in landslide susceptibility modeling,even when using the same ML algorithm.This research aims to compare the combinations of inventory data handling,cross validation(CV),and hyperparameter tuning strategies to generate landslide susceptibility maps.The results are expected to provide a general strategy for landslide susceptibility modeling using ML techniques.The authors employed eight landslide inventory data handling scenarios to convert a landslide polygon into a landslide point,i.e.,the landslide point is located on the toe(minimum height),on the scarp(maximum height),at the center of the landslide,randomly inside the polygon(1 point),randomly inside the polygon(3 points),randomly inside the polygon(5 points),randomly inside the polygon(10 points),and 15 m grid sampling.Random forest models using CV-nonspatial hyperparameter tuning,spatial CV-spatial hyperparameter tuning,and spatial CV-forward feature selection-no hyperparameter tuning were applied for each data handling strategy.The combination generated 24 random forest ML workflows,which are applied using a complete inventory of 743 landslides triggered by Tropical Cyclone Cempaka(2017)in Pacitan Regency,Indonesia,and 11 landslide controlling factors.The results show that grid sampling with spatial CV and spatial hyperparameter tuning is favorable because the strategy can minimize overfitting,generate a relatively high-performance predictive model,and reduce the appearance of susceptibility artifacts in the landslide area.Careful data inventory handling,CV,and hyperparameter tuning strategies should be considered in landslide susceptibility modeling to increase the applicability of landslide susceptibility maps in practical application.展开更多
As the risk of adverse impacts from tropical cyclones increases globally,more people and assets are being exposed to the potential damage they may cause.However,not much is known about the processes of flash floods tr...As the risk of adverse impacts from tropical cyclones increases globally,more people and assets are being exposed to the potential damage they may cause.However,not much is known about the processes of flash floods triggered by tropical cyclones,evacuation,and post-disaster recovery,particularly in Indonesia’s coastal regions.This study focuses on the coastal village of Wowong,located in Lembata Island,East Nusa Tenggara Province,which was affected by flash floods caused by the 2021 Tropical Cyclone Seroja.The objectives are:(1)to simulate flash floods based on extreme rainfall conditions during the cyclone and(2)to analyze the post-disaster recovery conditions following the flash floods.The method used to simulate flash floods is unsteady flow through the HEC-RAS hydraulic model,while post-disaster recovery analysis is conducted through focus group discussions and household surveys of flood-affected communities.The simulation shows that almost the entire village of Wowong was inundated with floodwaters,with a maximum depth reaching 3 m.The flood model was validated through statistical tests,producing a Kendall-Tau correlation coefficient of 0.59(p=2.2e-08).The sectors investigated for their recovery levels include water and sanitation,housing and infrastructure,livelihoods,as well as education and health.Recovery mechanisms largely stem from community self-help organizations,which are considered an uninstitutionalized social capital,as well as support from the government and non-governmental organizations.This study can serve as a consideration for local governments in prioritizing development and integrating social networks in areas affected by flash floods to prepare for disaster anticipatory action.展开更多
基金funded by the Faculty of Geography under the scheme of“Dana Hibah Penelitian Mandiri Dosen Tahun 2023 Tahap 1”。
文摘Ten rock samples consisting of one pyroclastic density current(PDC1)deposit,seven lava flows(LF1–7),and two summit lava domes(LD1,2)were studied to understand the petrogenesis and magma dynamics at Mt.Sumbing.The stratigraphy is arranged as LF1,PDC1,LF2,LF3,LF4,LF5,LF6,LF7,LD1,and LD2;furthermore,these rocks were divided into two types.TypeⅠ,observed in the oldest(LF1)sample,has poor MgO and high Ba/Nb,Th/Yb and Sr.The remaining samples(PDC1–LD2)represent typeⅡ,characterized by high MgO and low Ba/Nb,Th/Yb and Sr values.We suggest that type I is derived from AOC(altered oceanic crust)-rich melts that underwent significant crustal assimilation,while typeⅡoriginates from mantle-rich melts with less significant crustal assimilation.The early stage of typeⅡmagma(PDC1–LF3)was considered a closed system,evolving basaltic andesite into andesite(55.0–60.2 wt%SiO_(2))with a progressively increasing phenocryst(0.30–0.48φ_(PC))and decreasing crystal size distribution(CSD)slope(from-3.9 to-2.9).The evidence of fluctuating silica and phenocryst contents(between 55.9–59.7 wt%and 0.25–0.41φ_(PC),respectively),coupled with the kinked and steep(from-5.0 to-3.3)CSD curves imply the interchanging condition between open(i.e.,magma mixing)and closed magmatic systems during the middle stage(LF4–LF6).Finally,it underwent to closed system again during the final stage(LF7–LD2)because the magma reached dacitic composition(at most 68.9 wt%SiO_(2))with abundant phenocryst(0.38–0.45φ_(PC))and gentle CSD slope(from-4.1 to-1.2).
基金financially supported by the National Natural Science Foundation of China(Grant No.31070400)National Basic Research Program of China(Grant No.2010CB833501-01)Innovation Project of the Institute of Geographic Sciences and Natural Resources Research,CAS(Grant No.201003001)
文摘High concentration ground-level ozone(O3)has adverse effects on plant growth and photosynthesis.Compared to the O3concentration-based index,the O3flux-based(especially stomatal O3uptake)index has been considered the better criterion for assessing the impact of ozone on vegetation and ecosystems.This paper reports on a study of O3flux using the eddy covariance technique over a corn field in the Northwestern Shandong Plain of China.Diurnal variation of atmospheric O3concentration,deposition velocity and flux,and their relationships to environmental factors are analyzed.The results show that:(1)During the observation period(9 August–28 September,2011),there was a strong diurnal variation of O3concentration,with low(16.5 nL L?1)and high(60.1 nL L?1)O3mean concentrations observed around 6:30 and 16:00,respectively.Mean O3concentrations during daytime(6:00–18:00)and nighttime(18:00–6:00)were 39.8±23.1 and 20.7±14.1 nL L?1(mean±std),respectively.The maximum observed concentration was 97.5 nL L?1.The concentration was mainly affected by solar radiation and air temperature.(2)Whether daytime or nighttime,ground-level O3flux is always downward.The diurnal course of mean deposition velocity was divided into 4 phases:a low and stable process during nighttime,fast increasing in early morning,relatively large and steady changes around noon,and quickly decreasing in later afternoon.Daytime and nighttime mean deposition velocities were 0.29 and 0.09 cm s?1,respectively.The maximum deposition velocity was 0.81 cm s?1.The magnitude of deposition velocity was influenced by the corn growth period,and its diurnal variation was significantly correlated with global radiation and relative humidity.(3)O3flux was affected by variations of both O3concentration and deposition velocity,with mean O3fluxes-317.7 and?70.2 ng m?2s?1during daytime and nighttime,respectively.There was strong correlation between O3flux and CO2flux or latent heat flux.By comparing the deposition velocities of daytime and nighttime,we infer that stomatal uptake was probably the main sink of ground-level O3.
文摘Machine learning(ML)algorithms are frequently used in landslide susceptibility modeling.Different data handling strategies may generate variations in landslide susceptibility modeling,even when using the same ML algorithm.This research aims to compare the combinations of inventory data handling,cross validation(CV),and hyperparameter tuning strategies to generate landslide susceptibility maps.The results are expected to provide a general strategy for landslide susceptibility modeling using ML techniques.The authors employed eight landslide inventory data handling scenarios to convert a landslide polygon into a landslide point,i.e.,the landslide point is located on the toe(minimum height),on the scarp(maximum height),at the center of the landslide,randomly inside the polygon(1 point),randomly inside the polygon(3 points),randomly inside the polygon(5 points),randomly inside the polygon(10 points),and 15 m grid sampling.Random forest models using CV-nonspatial hyperparameter tuning,spatial CV-spatial hyperparameter tuning,and spatial CV-forward feature selection-no hyperparameter tuning were applied for each data handling strategy.The combination generated 24 random forest ML workflows,which are applied using a complete inventory of 743 landslides triggered by Tropical Cyclone Cempaka(2017)in Pacitan Regency,Indonesia,and 11 landslide controlling factors.The results show that grid sampling with spatial CV and spatial hyperparameter tuning is favorable because the strategy can minimize overfitting,generate a relatively high-performance predictive model,and reduce the appearance of susceptibility artifacts in the landslide area.Careful data inventory handling,CV,and hyperparameter tuning strategies should be considered in landslide susceptibility modeling to increase the applicability of landslide susceptibility maps in practical application.
文摘As the risk of adverse impacts from tropical cyclones increases globally,more people and assets are being exposed to the potential damage they may cause.However,not much is known about the processes of flash floods triggered by tropical cyclones,evacuation,and post-disaster recovery,particularly in Indonesia’s coastal regions.This study focuses on the coastal village of Wowong,located in Lembata Island,East Nusa Tenggara Province,which was affected by flash floods caused by the 2021 Tropical Cyclone Seroja.The objectives are:(1)to simulate flash floods based on extreme rainfall conditions during the cyclone and(2)to analyze the post-disaster recovery conditions following the flash floods.The method used to simulate flash floods is unsteady flow through the HEC-RAS hydraulic model,while post-disaster recovery analysis is conducted through focus group discussions and household surveys of flood-affected communities.The simulation shows that almost the entire village of Wowong was inundated with floodwaters,with a maximum depth reaching 3 m.The flood model was validated through statistical tests,producing a Kendall-Tau correlation coefficient of 0.59(p=2.2e-08).The sectors investigated for their recovery levels include water and sanitation,housing and infrastructure,livelihoods,as well as education and health.Recovery mechanisms largely stem from community self-help organizations,which are considered an uninstitutionalized social capital,as well as support from the government and non-governmental organizations.This study can serve as a consideration for local governments in prioritizing development and integrating social networks in areas affected by flash floods to prepare for disaster anticipatory action.
