A Global Land Cover Climatology Using MODIS Data Broxton, Patrick D.; Zeng, Xubin; Sulla-Menashe, Damien ...
Journal of applied meteorology and climatology,
06/2014, Letnik:
53, Številka:
6
Journal Article
Recenzirano
Odprti dostop
Global land cover data are widely used in weather, climate, and hydrometeorological models. The Collection 5.1 Moderate Resolution Imaging Spectroradiometer (MODIS) Land Cover Type (MCD12Q1) product ...is found to have a substantial amount of interannual variability, with 40% of land pixels showing land cover change one or more times during 2001–10. This affects the global distribution of vegetation if any one year or many years of data are used, for example, to parameterize land processes in regional and global models. In this paper, a value-added global 0.5-km land cover climatology (a single representative map for 2001–10) is developed by weighting each land cover type by its corresponding confidence score for each year and using the highest-weighted land cover type in each pixel in the 2001–10 MODIS data. The climatology is validated by comparing it with the System for Terrestrial Ecosystem Parameterization database as well as additional pixels that are identified from the Google Earth proprietary software database. When compared with the data of any individual year, this climatology does not substantially alter the overall global frequencies of most land cover classes but does affect the global distribution of many land cover classes. In addition, it is validated as well as or better than the MODIS data for individual years. Also, it is based on higher-quality data and is validated better than the Global Land Cover Characteristics database, which is based on 1 year of Advanced Very High Resolution Radiometer data and represents a widely used first-generation global product.
Celotno besedilo
Dostopno za:
BFBNIB, DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
Commonly used bias correction methods such as quantile mapping (QM) assume the function of error correction values between modeled and observed distributions are stationary or time invariant. This ...article finds that this function of the error correction values cannot be assumed to be stationary. As a result, QM lacks justification to inflate/deflate various moments of the climate change signal. Previous adaptations of QM, most notably quantile delta mapping (QDM), have been developed that do not rely on this assumption of stationarity. Here, we outline a methodology called scaled distribution mapping (SDM), which is conceptually similar to QDM, but more explicitly accounts for the frequency of rain days and the likelihood of individual events. The SDM method is found to outperform QM, QDM, and detrended QM in its ability to better preserve raw climate model projected changes to meteorological variables such as temperature and precipitation.
Time‐variable transit time distributions (TTDs) have been utilized as a tool to understand how catchments transmit water. However, most of the existing TTD estimation methods require to impose ...certain structures on those TTDs a priori, which could lead to misinterpreting data. We present a data‐based method to estimate time‐variable TTDs without imposing their structure a priori. The core of the method is the use of a revised flow‐weighted time, where TTDs do not reflect variable external forcings directly. The functional forms of the TTDs are much simpler in flow‐weighted time, compared to those in calendar time, and this allows for easier estimation of TTDs. Dynamic (state‐dependent) multiple linear regression methods were applied to estimate the time‐variable TTDs in flow‐weighted time, which can eventually be transformed back to calendar time. The method performs well in a proof‐of‐concept demonstration with synthetic data sets. We also discuss potential generalizations of the proposed method.
Key Points
We outline a data‐based method to estimate transit time distributions (TTDs) without imposing their structure a priori
We use a (dynamic) multiple linear regression method to estimate (time‐variable) TTDs in flow‐weighted time
In a proof‐of‐concept demonstration, the method estimates TTDs close to the actual TTDs
Flow recession analysis, relating discharge Q and its time rate of change −dQ/dt, has been widely used to understand catchment scale flow dynamics. However, data points in the recession plot, the ...plot of −dQ/dt versus Q, typically form a wide point cloud due to noise and hysteresis in the storage‐discharge relationship, and it is still unclear what information we can extract from the plot and how to understand the information. There seem to be two contrasting approaches to interpret the plot. One emphasizes the importance of the ensemble characteristics of many recessions (i.e., the lower envelope or a measure of central tendency), and the other highlights the importance of the event scale analysis and questions the meaning of the ensemble characteristics. We examine if those approaches can be reconciled. We utilize a machine learning tool to capture the point cloud using the past trajectory of daily discharge. Our model results for a catchment show that most of the data points can be captured using 5 days of past discharge. We show that we can learn the catchment scale flow recession dynamics from what the machine learned. We analyze patterns learned by the machine and explain and hypothesize why the machine learned those characteristics. The hysteresis in the plot mainly occurs during the early time dynamics, and the flow recession dynamics eventually converge to an attractor in the plot, which represents the master recession curve. We also illustrate that a hysteretic storage‐discharge relationship can be estimated based on the attractor.
Key Points
A machine learning tool captures time‐variable flow recession dynamics that identify scanning curves of the storage‐discharge relationship
Machine learned individual flow recession curves converge to a common attractor in the recession plot, revealing the master recession curve
It leads to a novel way of analyzing the recession plot, unifying the event‐based analysis and the analysis of ensemble characteristics
We observed water fluxes and isotopic compositions within the subsurface of six small nested zero‐order catchments over the course of three North American monsoon seasons and found that mean transit ...times (mTTs) were variable between seasons and different spatial patterns of mTTs emerged each year. For each monsoon season, it was possible to correlate mTTs with a different physical catchment property. In 2007, mTTs correlated best with mean soil depth, in 2008 soil hydraulic conductivity gained importance in explaining the variability and in 2009 planform curvature showed the best correlation. Differences in meteorological forcing between the three monsoon seasons explained the temporal variability of mTTs. In 2007, a series of precipitation events caused the storage capacity of the soils of some of the zero‐order catchments to be exceeded. As a result those catchments started producing quick runoff (overland and macropore flow). In 2008, precipitation events were more evenly distributed throughout the season, soils did not saturate, runoff coefficients decreased because more water left the catchment via evapotranspiration and soil hydraulic conductivity became a stronger control since matrix flow dominated. The 2009 monsoon was unusually dry, the soil storage became depleted and water flowed mainly through bedrock pathways. Therefore, topographic parameters gained importance in determining how quickly water arrived at the catchment outlet. In order to improve our understanding of what controls mTTs we suggest a dimensionless number that helps identifying partitioning thresholds and sorts precipitation events into one of the three response modes that were observed in our zero‐order catchments.
Key Points
Inherent catchment properties and external forcings control transit times
External forcings sort precipitation events into specific response domains
Inherent catchment properties control transit times within these domains
Large-scale biogeographical shifts in vegetation are predicted in response to the altered precipitation and temperature regimes associated with global climate change. Vegetation shifts have profound ...ecological impacts and are an important climate-ecosystem feedback through their alteration of carbon, water, and energy exchanges of the land surface. Of particular concern is the potential for warmer temperatures to compound the effects of increasingly severe droughts by triggering widespread vegetation shifts via woody plant mortality. The sensitivity of tree mortality to temperature is dependent on which of 2 non-mutually-exclusive mechanisms predominates--temperature-sensitive carbon starvation in response to a period of protracted water stress or temperature-insensitive sudden hydraulic failure under extreme water stress (cavitation). Here we show that experimentally induced warmer temperatures (almost equal to4 °C) shortened the time to drought-induced mortality in Pinus edulis (piñon shortened pine) trees by nearly a third, with temperature-dependent differences in cumulative respiration costs implicating carbon starvation as the primary mechanism of mortality. Extrapolating this temperature effect to the historic frequency of water deficit in the southwestern United States predicts a 5-fold increase in the frequency of regional-scale tree die-off events for this species due to temperature alone. Projected increases in drought frequency due to changes in precipitation and increases in stress from biotic agents (e.g., bark beetles) would further exacerbate mortality. Our results demonstrate the mechanism by which warmer temperatures have exacerbated recent regional die-off events and background mortality rates. Because of pervasive projected increases in temperature, our results portend widespread increases in the extent and frequency of vegetation die-off.
Recent field observations indicate that in many forest ecosystems, plants use water that may be isotopically distinct from soil water that ultimately contributes to streamflow. Such an assertion has ...been met with varied reactions. Of the outstanding questions, we examine whether ecohydrological separation of water between trees and streams results from a separation in time, or in space. Here we present results from a 9‐month drought and rewetting experiment at the 26,700‐m3 mesocosm, Biosphere 2‐Tropical Rainforest biome. We test the null hypothesis that transpiration and groundwater recharge water are sampled from the same soil volume without preference for old nor young water. After a 10‐week drought, we added 66 mm of labeled rainfall with 152‰ δ2H distributed over four events, followed by background rainfall (−60‰ δ2H) distributed over 13 events. Our results show that mean transit times through groundwater recharge and plant transpiration were markedly different: groundwater recharge was 2–7 times faster (~9 days) than transpired water (range 17–62 days). The “age” of transpired water showed strong dependence on species and was linked to the difference between midday leaf water potential and soil matric potential. Moreover, our results show that trees used soil water (89% ±6) and not the “more mobile” (represented by “zero tension” seepage) water (11% ±6). The finding, which rejects our null hypothesis, is novel in that this partitioning is established based on soil water residence times. Our study quantifies mean transit times for transpiration and seepage flows under dynamic conditions.
Plain Language Summary
Recent studies suggest that plants use a type of water that is different to the water that recharges the ground, a phenomenon described as the two water worlds. It is unclear, however, whether these waters are segregated in space or in time. That is, do plants draw water from parts of the soil different to groundwater recharge, or do plant water withdrawals happen at a different time from groundwater recharge? Evidence from well‐controlled experiments is badly needed because the two water worlds, if true, means that our understanding of the water cycle is incomplete. Here we perform a 9‐month drought and rainfall experiment, taking fingerprints of the water molecule, to follow a raindrop from the moment it enters the ground through to its exit via plants or groundwater recharge. Results point to two main discoveries: (1) the travel time of water via root water uptake is much longer than the travel time of water leading to groundwater recharge and (2) the water taken by tree roots comes from parts of the soil that are different to the water leading to groundwater recharge. These discoveries show the segregation of these two components of the water cycle in space and in time.
Key Points
Root water uptake is derived from the less mobile water in the soil matrix, different to the more mobile water component in soils. The transit times (“ages”) of water taken by roots are older than seepage (“groundwater recharge”) water by a factor of 2 to 7
Ecohydrological separation suggests that time‐sensitive sampling and modeling techniques are critical for understanding the water cycle
Species‐specific differences in root water uptake transit times suggest that trees should not be treated as simple transport vessels (or “straws”) in land surface models
Global land-cover data are widely used in regional and global models because land cover influences land–atmosphere exchanges of water, energy, momentum, and carbon. Many models use data of maximum ...green vegetation fraction (MGVF) to describe vegetation abundance. MGVF products have been created in the past using different methods, but their validation with ground sites is difficult. Furthermore, uncertainty is introduced because many products use a single year of satellite data. In this study, a global 1-km MGVF product is developed on the basis of a “climatology” of data of Moderate Resolution Imaging Spectroradiometer (MODIS) normalized difference vegetation index and land-cover type, which removes biases associated with unusual greenness and inaccurate land-cover classification for individual years. MGVF shows maximum annual variability from 2001 to 2012 for intermediate values of average MGVF, and the standard deviation of MGVF normalized by its mean value decreases nearly monotonically as MGVF increases. In addition, there are substantial differences between this climatology and MGVF data from the MODIS Continuous Fields (CF) Collection 3, which is currently used in the Community Land Model. Although the CF data only use 2001 MODIS data, many of these differences cannot be explained by usage of different years of data. In particular, MGVF as based on CF data is usually higher than that based on the MODIS climatology from this paper. It is difficult to judge which product is more realistic because of a lack of ground truth, but this new MGVF product is more consistent than the CF data with the MODIS leaf area index product (which is also used to describe vegetation abundance in models).
Celotno besedilo
Dostopno za:
BFBNIB, DOBA, IZUM, KILJ, NUK, PILJ, PNG, SAZU, SIK, UILJ, UKNU, UL, UM, UPUK
The transit time of water is an important indicator of catchment functioning and affects many biological and geochemical processes. Water entering a catchment at one point in time is composed of ...water molecules that will spend different amounts of time in the catchment before exiting. The next water input pulse can exhibit a totally different distribution of transit times. The distribution of water transit times is thus best characterized by a time‐variable probability density function. It is often assumed, however, that the variability of the transit time distribution is negligible and that catchments can be characterized with a unique transit time distribution. In many cases this assumption is not valid because of variations in precipitation, evapotranspiration, and catchment water storage and associated (de)activation of dominant flow paths. This paper presents a general method to estimate the time‐variable transit time distribution of catchment waters. Application of the method using several years of rainfall‐runoff and stable water isotope data yields an ensemble of transit time distributions with different moments. The combined probability density function represents the master transit time distribution and characterizes the intra‐annual and interannual variability of catchment storage and flow paths. Comparing the derived master transit time distributions of two research catchments (one humid and one semiarid) reveals differences in dominant hydrologic processes and dynamic water storage behavior, with the semiarid catchment generally reacting slower to precipitation events and containing a lower fraction of preevent water in the immediate hydrologic response.
Key Points
Water transit time distributions are highly irregular and variable in time
Water transit time distributions differ from hydrologic response functions
Differences between the two functions yield information on storage dynamics
The notion that landscape features have coevolved over time is well known in the Earth sciences. Hydrologists have recently called for a more rigorous connection between emerging spatial patterns of ...landscape features and the hydrological response of catchments, and have termed this concept catchment coevolution. In this paper we review recent literature on this subject and attempt to synthesize what we have learned into a general framework that would improve predictions of hydrologic change. We first present empirical evidence of the interaction and feedback of landscape evolution and changes in hydrological response. From this review it is clear that the independent drivers of catchment coevolution are climate, geology, and tectonics. We identify common currency that allows comparing the levels of activity of these independent drivers, such that, at least conceptually, we can quantify the rate of evolution or aging. Knowing the hydrologic age of a catchment by itself is not very meaningful without linking age to hydrologic response. Two avenues of investigation have been used to understand the relationship between (differences in) age and hydrological response: (i) one that is based on relating present landscape features to runoff processes that are hypothesized to be responsible for the current fingerprints in the landscape; and (ii) one that takes advantage of an experimental design known as space‐for‐time substitution. Both methods have yielded significant insights in the hydrologic response of landscapes with different histories. If we want to make accurate predictions of hydrologic change, we will also need to be able to predict how the catchment will further coevolve in association with changes in the activity levels of the drivers (e.g., climate). There is ample evidence in the literature that suggests that whole‐system prediction of catchment coevolution is, at least in principle, plausible. With this imperative we outline a research agenda that implements the concepts of catchment coevolution for building a holistic framework toward improving predictions of hydrologic change.
Key Points:
Catchments coevolve in function of climate, geology, and tectonics
This coevolution leads to spatial patterns of landscape features
These landscape features can inform models to predict hydrologic change
PDF
