Source code for landlab.components.lake_fill.lake_fill_barnes

#!/usr/env/python

"""lake_fill_barnes.py.

Fill sinks in a landscape to the brim, following the Barnes et al.
(2014) algorithms.
"""


import heapq
import itertools
from collections import deque

import numpy as np

from landlab import Component
from landlab import NodeStatus
from landlab import RasterModelGrid
from landlab.components import FlowAccumulator
from landlab.utils import StablePriorityQueue
from landlab.utils.return_array import return_array_at_node

LARGE_ELEV = 9999999999.0

# TODO: Needs to have rerouting functionality...


def _fill_one_node_to_flat(fill_surface, all_neighbors, pitq, openq, closedq, dummy):
    """Implements the Barnes et al. algorithm for a simple fill. Assumes the
    _open and _closed lists have already been updated per Barnes algos 2&3, lns
    1-7.

    Parameters
    ----------
    fill_surface : 1-D array of length nnodes
        The surface to fill in LL node order. Modified in place.
    all_neighbors : (nnodes, max_nneighbours) array
        Adjacent nodes at each node.
    pitq : heap queue (i.e., a structured list)
        Current nodes known to be in a lake, if already identified.
    openq : StablePriorityQueue object
        Ordered queue of nodes remaining to be checked out by the algorithm
        that are known not to be in a lake.
    closedq : 1-D boolean array of length nnodes
        Nodes already or not to be explored by the algorithm.
    dummy : any Python object
        Necessary for direct comparison with _fill_one_node_to_slant.

    Examples
    --------
    >>> import numpy as np
    >>> from landlab import RasterModelGrid
    >>> mg = RasterModelGrid((5, 6))
    >>> for edge in ("left", "top", "bottom"):
    ...     mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED
    ...
    >>> z = np.array(
    ...     [
    ...         [0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
    ...         [0.0, 2.1, 1.1, 0.6, 1.6, 0.0],
    ...         [0.0, 2.0, 1.0, 0.5, 1.5, 0.0],
    ...         [0.0, 2.2, 1.2, 0.7, 1.7, 0.0],
    ...         [0.0, 0.0, 0.0, 0.0, 0.0, 0.0],
    ...     ]
    ... ).flatten()
    >>> zw = z.copy()
    >>> openq = StablePriorityQueue()
    >>> pitq = []
    >>> closedq = mg.zeros("node", dtype=bool)
    >>> closedq[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True
    >>> edges = np.array([11, 17, 23])
    >>> for edgenode in edges:
    ...     openq.add_task(edgenode, priority=z[edgenode])
    ...
    >>> closedq[edges] = True
    >>> while True:
    ...     try:
    ...         _fill_one_node_to_flat(
    ...             zw, mg.adjacent_nodes_at_node, pitq, openq, closedq, None
    ...         )
    ...     except KeyError:
    ...         break
    ...

    Now check the values make sense.

    >>> lake = np.array(
    ...     [
    ...         [0, 0, 0, 0, 0, 0],
    ...         [0, 0, 1, 1, 0, 0],
    ...         [0, 0, 1, 1, 0, 0],
    ...         [0, 0, 1, 1, 0, 0],
    ...         [0, 0, 0, 0, 0, 0],
    ...     ],
    ...     dtype=bool,
    ... ).flatten()

    >>> np.allclose(zw[lake], z[16])
    True
    >>> np.all(np.greater(zw[lake], z[lake]))
    True
    >>> np.allclose(zw[~lake], z[~lake])
    True
    """
    try:
        c = heapq.heappop(pitq)
    except IndexError:
        c = openq.pop_task()
        # this will raise a KeyError once it's exhausted both queues
    cneighbors = all_neighbors[c]
    openneighbors = cneighbors[np.logical_not(closedq[cneighbors])]  # for efficiency
    closedq[openneighbors] = True
    for n in openneighbors:
        if fill_surface[n] <= fill_surface[c]:
            fill_surface[n] = fill_surface[c]
            heapq.heappush(pitq, n)
        else:
            openq.add_task(n, priority=fill_surface[n])


[docs] class LakeMapperBarnes(Component): """A Landlab implementation of the Barnes et al. (2014) lake filling & lake routing algorithms, lightly modified and adapted for Landlab by DEJH. This component is designed as a direct replacement for the LakeMapper as existing pre-Aug 2018, and provides a suite of properties to access information about the lakes created each time it is run. Only significant difference is the way the lakes are coded: this component uses the (unique) ID of the outlet node, whereas DepressionFinderAndRouter uses one of the pit node IDs. Note also this component does not offer the `lake_codes` or `display_depression_map` options, for essentially this reason. Use `lake_map` instead for both. It also uses a much more Landlabbian `run_one_step()` method as its driver, superceding DepressionFinderAndRouter's `map_depressions()`. A variety of options is provided. Flow routing is route-to-one in this implementation, but can be either D4 ("steepest") or D8 on a raster. The surface can be filled to either flat or a very slight downward incline, such that subsequent flow routing will run over the lake surface. This incline is applied at machine precision to minimise the chances of creating false outlets by overfill, and note that the gradient as calculated on such surfaces may still appear to be zero. The filling can either be performed in place, or on a new (water) surface distinct from the original (rock) surface. For efficiency, data structures describing the lakes and their properties are only created, and existing flow direction and flow accumulation fields modified, if those flags are set at instantiation. With care, this component can be used to create a dynamically flooding surface in a fluvial landscape (interacting with, e.g., the StreamPowerEroder). See the run_one_step docstring for an example. References ---------- **Required Software Citation(s) Specific to this Component** Barnes, R., Lehman, C., Mulla, D. (2014). Priority-flood: An optimal depression-filling and watershed-labeling algorithm for digital elevation models. Computers and Geosciences 62(C), 117 - 127. https://dx.doi.org/10.1016/j.cageo.2013.04.024 **Additional References** None Listed """ _name = "LakeMapperBarnes" _cite_as = """ @article{BARNES2014117, title = {Priority-flood: An optimal depression-filling and watershed-labeling algorithm for digital elevation models}, journal = "Computers & Geosciences", volume = "62", pages = "117 - 127", year = "2014", issn = "0098-3004", doi = "https://doi.org/10.1016/j.cageo.2013.04.024", url = "http://www.sciencedirect.com/science/article/pii/S0098300413001337", author = "Richard Barnes and Clarence Lehman and David Mulla", keywords = "Pit filling, Terrain analysis, Hydrology, Drainage network, Modeling, GIS" }""" _unit_agnostic = True _info = { "drainage_area": { "dtype": float, "intent": "inout", "optional": False, "units": "m**2", "mapping": "node", "doc": "Upstream accumulated surface area contributing to the node's discharge", }, "flow__data_structure_delta": { "dtype": int, "intent": "inout", "optional": False, "units": "-", "mapping": "node", "doc": ( "Node array containing the elements delta[1:] of the data " "structure 'delta' used for construction of the " "downstream-to-upstream node array" ), }, "flow__link_to_receiver_node": { "dtype": int, "intent": "inout", "optional": False, "units": "-", "mapping": "node", "doc": "ID of link downstream of each node, which carries the discharge", }, "flow__receiver_node": { "dtype": int, "intent": "inout", "optional": False, "units": "-", "mapping": "node", "doc": "Node array of receivers (node that receives flow from current node)", }, "flow__sink_flag": { "dtype": bool, "intent": "inout", "optional": False, "units": "-", "mapping": "node", "doc": "Boolean array, True at local lows", }, "flow__upstream_node_order": { "dtype": int, "intent": "inout", "optional": False, "units": "-", "mapping": "node", "doc": "Node array containing downstream-to-upstream ordered list of node IDs", }, "surface_water__discharge": { "dtype": float, "intent": "inout", "optional": False, "units": "m**3/s", "mapping": "node", "doc": "Volumetric discharge of surface water", }, "topographic__elevation": { "dtype": float, "intent": "inout", "optional": False, "units": "m", "mapping": "node", "doc": "Land surface topographic elevation", }, }
[docs] def __init__( self, grid, surface="topographic__elevation", method="Steepest", fill_flat=True, fill_surface="topographic__elevation", redirect_flow_steepest_descent=False, reaccumulate_flow=False, ignore_overfill=False, track_lakes=True, ): """Initialize the component. Parameters ---------- grid : ModelGrid A grid. surface : field name at node or array of length node The surface to direct flow across. method : {'Steepest', 'D8'} Whether or not to recognise diagonals as valid flow paths, if a raster. Otherwise, no effect. fill_flat : bool If True, pits will be filled to perfectly horizontal. If False, the new surface will be slightly inclined to give steepest descent flow paths to the outlet. fill_surface : bool Sets the field or array to fill. If fill_surface is surface, this operation occurs in place, and is faster. Note that the component will overwrite fill_surface if it exists; to supply an existing water level to it, supply that water level field as surface, not fill_surface. redirect_flow_steepest_descent : bool If True, the component outputs modified versions of the 'flow__receiver_node', 'flow__link_to_receiver_node', 'flow__sink_flag', and 'topographic__steepest_slope' fields. These are the fields output by the FlowDirector components, so set to True if you wish to pass this LakeFiller to the FlowAccumulator, or if you wish to work directly with the new, correct flow directions and slopes without rerunning these components on your new surface. Ensure the necessary fields already exist, and have already been calculated by a FlowDirector! This also means you need to instantiate your FlowDirector **before** you instantiate the LakeMapperBarnes. Note that the new topographic__steepest_slope will always be set to zero, even if fill_flat=False (i.e., there is actually a miniscule gradient on the new topography, which gets ignored). reaccumulate_flow : bool If True, and redirect_flow_steepest_descent is True, the run method will (re-)accumulate the flow after redirecting the flow. This means the 'drainage_area', 'surface_water__discharge', 'flow__upstream_node_order', and the other various flow accumulation fields (see output field names) will now reflect the new drainage patterns without having to manually reaccumulate the discharge. If True but redirect_flow_steepest_descent is False, raises an ValueError. ignore_overfill : bool If True, suppresses the Error that would normally be raised during creation of a gentle incline on a fill surface (i.e., if not fill_flat). Typically this would happen on a synthetic DEM where more than one outlet is possible at the same elevation. If True, the was_there_overfill property can still be used to see if this has occurred. track_lakes : bool If True, the component permits a slight hit to performance in order to explicitly track which nodes have been filled, and to enable queries on that data in retrospect. Set to False to simply fill the surface and be done with it. """ super().__init__(grid) if "flow__receiver_node" in grid.at_node and grid.at_node[ "flow__receiver_node" ].size != grid.size("node"): raise NotImplementedError( "A route-to-multiple flow director has been " "run on this grid. The landlab development team has not " "verified that LakeMapperBarnes is compatible with " "route-to-multiple methods. Please open a GitHub Issue " "to start this process." ) self._pit = [] self._closed = self._grid.zeros(at="node", dtype=bool) self._gridclosednodes = ( self._grid.status_at_node == self._grid.BC_NODE_IS_CLOSED ) # close up the BC_NODE_IS_CLOSED permanently: self._closed[self._gridclosednodes] = True # this component maintains its own internal count of how often it has # been called. This is to enable "cheap" data access of the various # available data structures without needless recalculation self._runcounter = itertools.count() self._runcount = -1 # not yet run self._lastcountforlakemap = -1 # lake_map has not yet been called self._PitTop = LARGE_ELEV # variable to not overfill slanted surfaces self._ignore_overfill = ignore_overfill self._overfill_flag = False self._track_lakes = track_lakes # get the neighbour call set up: if method not in {"Steepest", "D8"}: raise ValueError(f"{method}: method must be 'Steepest' or 'D8'") if method == "D8": if isinstance(grid, RasterModelGrid): self._allneighbors = np.concatenate( ( self._grid.adjacent_nodes_at_node, self._grid.diagonal_adjacent_nodes_at_node, ), axis=1, ) else: # not a raster raise ValueError( ( "D8 is not a valid value for method if grid type is " + "{gridtype}!" ).format(gridtype=type(grid)) ) else: self._allneighbors = self._grid.adjacent_nodes_at_node # A key difference from the "pure" Barnes algorithm for LL is that # we must'n flood from all the edges. Instead, we can only flood from # a true grid edge, i.e., only the FIXED boundary types. (Both # CLOSED and LOOPED assume flow either can't get out there, or at # least, there's more land in that direction that will be handled # otherwise.) Note we add a test that there is at least some kind # of outlet!! self._edges = np.where( np.logical_or( self._grid.status_at_node == NodeStatus.FIXED_VALUE, self._grid.status_at_node == NodeStatus.FIXED_GRADIENT, ) )[0] if self._edges.size == 0: raise ValueError( "No valid outlets found on the grid! You cannot run the " + "filling algorithms!" ) # and finally, close these up permanently as well (edges will always # be edges...) self._closed[self._edges] = True # ...note there's a slight of hand here, Because of the ordering of LL # grids, the last node will always be a boundary node, even for very # odd Voronois. This enables us to treat out -1s in the neighbour # arrays as always True. But, just in case... assert self._closed[-1] # check if we are modifying in place or not. This gets used to check # it makes sense to calculate various properties. self._inplace = surface is fill_surface # then self._surface = return_array_at_node(grid, surface) self._fill_surface = return_array_at_node(grid, fill_surface) # NOTE: buggy functionality of return_array_at_node here means # component can't yet handle arrays as opposed to fields... # This will be resolved by a modification to return_array_at_node # now, work out what constitutes a "surface" under various input opts: self._dontredirect = not redirect_flow_steepest_descent if redirect_flow_steepest_descent: # this routine only permitted if we store the lake info, so if not track_lakes: raise ValueError("You must track_lakes to redirect the flow!") # Check we have the necessary fields already existing. # These will raise FieldErrors if they don't. # This will cause a bunch of our tests to break, so users will # never see this. self._receivers = self._grid.at_node["flow__receiver_node"] self._receiverlinks = self._grid.at_node["flow__link_to_receiver_node"] self._steepestslopes = self._grid.at_node["topographic__steepest_slope"] # if raster, do the neighbors & diagonals separate when rerouting # so we'll need to pull these separately: if method == "D8": # Raster test unnecessary given tests above self._neighbor_arrays = ( self._grid.adjacent_nodes_at_node, self._grid.diagonal_adjacent_nodes_at_node, ) self._link_arrays = ( self._grid.links_at_node, self._grid.d8s_at_node[:, 4:], ) self._neighbor_lengths = self._grid.length_of_d8 else: self._neighbor_arrays = (self._grid.adjacent_nodes_at_node,) self._link_arrays = (self._grid.links_at_node,) self._neighbor_lengths = self._grid.length_of_link if reaccumulate_flow: if not redirect_flow_steepest_descent: raise ValueError( "You must also redirect_flow_steepest_descent if you " + "want to reaccumulate_flow!" ) self._reaccumulate = True self._fa = FlowAccumulator(self._grid, flow_director=method) else: self._reaccumulate = False self._fill_flat = fill_flat if fill_flat: self._fill_one_node = _fill_one_node_to_flat else: self._fill_one_node = self._fill_one_node_to_slant
def _fill_one_node_to_slant( self, fill_surface, all_neighbors, pitq, openq, closedq, ignore_overfill ): """Implements the Barnes et al. algorithm to obtain a naturally draining surface, updating a single node. Assumes the _open and _closed lists have already been updated per Barnes algos 2&3, lns 1-7. Parameters ---------- fill_surface : 1-D array of length nnodes The surface to fill in LL node order. Modified in place. all_neighbors : (nnodes, max_nneighbours) array Adjacent nodes at each node. pitq : heap queue (i.e., a structured list) Current nodes known to be in a lake, if already identified. openq : StablePriorityQueue object Ordered queue of nodes remaining to be checked out by the algorithm that are known not to be in a lake. closedq : 1-D boolean array of length nnodes Nodes already or not to be explored by the algorithm. ignore_overfill : bool If False, method will raise a ValueError if adding an increment to the node's elevation would fundamentally alter the resulting drainage pattern (e.g., it would create a new outlet somewhere). If True, the elevation of the node will be changed regardless. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> from landlab.utils import StablePriorityQueue >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z.reshape(mg.shape)[2, 1:-1] = [2.0, 1.0, 0.5, 1.5] >>> z.reshape(mg.shape)[1, 1:-1] = [2.1, 1.1, 0.6, 1.6] >>> z.reshape(mg.shape)[3, 1:-1] = [2.2, 1.2, 0.7, 1.7] >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> edges = np.array([11, 17, 23]) >>> open = StablePriorityQueue() >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> first_nodes_checked = [] >>> for i in range(3): # run a couple of steps ... lmb._fill_one_node_to_slant( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False ... ) ... print(open.peek_at_task()) ... assert lmb._pit == [] # these steps don't find pits ... 17 23 16 >>> lmb._fill_one_node_to_slant( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False ... ) >>> lmb._pit == [15] # Popping 16 off "open" puts 15 in "pit" True >>> np.isclose(z[15], z[16]) # 15 has also been filled in this step True >>> z[15] > z[16] # ...but 15 is incrementally greater than 16 True >>> lmb._fill_one_node_to_slant( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False ... ) >>> lmb._pit == [9, 21, 14] # 15 pops of pit, these neighbors go in True >>> np.allclose(z[15], [z[9], z[21], z[14]]) # now filled True >>> np.all([z[9] == z[21], z[21] == z[14]]) # these perfectly level True >>> z[9] > z[15] # ...but incrementally above 15 True >>> lmb._fill_one_node_to_slant( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False ... ) >>> lmb._pit == [8, 21, 14] # 9 popped off pit, 8 went in. And so on. True Test a failing example. This behaviour exists to prevent the application of the gradient from fundamentally altering the drainage pattern that "should" result. >>> mg = RasterModelGrid((3, 7)) >>> for edge in ("top", "right", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z.reshape(mg.shape)[1, 1:-1] = [1.0, 0.2, 0.1, 1.0000000000000004, 1.5] >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> open = StablePriorityQueue() >>> edges = np.array([7]) >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> while True: ... try: ... lmb._fill_one_node_to_slant( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False ... ) ... except KeyError: ... break ... except ValueError: ... print("ValueError was raised: here we overfilled") ... ValueError was raised: here we overfilled """ try: topopen = openq.peek_at_task() except KeyError: noopen = True else: noopen = False try: toppit = pitq[0] except IndexError: nopit = True else: nopit = False if ( not (nopit or noopen) and topopen == toppit ): # intentionally tight comparison # not clear how this occurs, but present in Barnes -> # DEJH suspects this should be an elevation comparison given the # text description. Regardless, this is only to ensure # repeatability, so it's not vital even if these cases don't # trigger c = openq.pop_task() self._PitTop = LARGE_ELEV if not nopit: c = heapq.heappop(pitq) if np.isclose(self._PitTop, LARGE_ELEV): self._PitTop = fill_surface[c] else: c = openq.pop_task() # again, returns KeyError if empty self._PitTop = LARGE_ELEV for n in all_neighbors[c]: if closedq[n]: continue else: closedq[n] = True nextval = np.nextafter(fill_surface[c], LARGE_ELEV) # DEJH believes that in the LL use cases this is impossible, # but retained as comments since present in Barnes algorithm # if self._gridclosednodes[n]: # heapq.heappush(pitq, n) if fill_surface[n] <= nextval: # former elif if self._PitTop < fill_surface[n] and nextval >= fill_surface[n]: if ignore_overfill: self._overfill_flag = True else: raise ValueError( "Pit is overfilled due to creation of two " + "outlets as the minimum gradient gets applied. " + "Suppress this Error with the ignore_overfill " + "flag at component instantiation." ) fill_surface[n] = nextval heapq.heappush(pitq, n) else: openq.add_task(n, priority=fill_surface[n]) def _fill_to_flat_with_tracking( self, fill_surface, all_neighbors, pitq, openq, closedq ): """Implements the Barnes et al. algorithm for a simple fill over the grid. Assumes the _open and _closed lists have already been updated per Barnes algos 2&3, lns 1-7. This version runs a little more slowly to enable tracking of which nodes are linked to which outlets. Parameters ---------- fill_surface : 1-D array The surface to fill in LL node order. Modified in place. all_neighbors : (nnodes, max_nneighbours) array Adjacent nodes at each node. pitq : heap queue (i.e., a structured list) Current nodes known to be in a lake, if already identified. openq : StablePriorityQueue object Ordered queue of nodes remaining to be checked out by the algorithm that are known not to be in a lake. closedq : 1-D boolean array of length nnodes Nodes already or not to be explored by the algorithm. Returns ------- lakemappings : {outlet_ID : [nodes draining to outlet]} Dict with outlet nodes of individual lakes as keys, and lists of each node inundated (i.e., depth > 0.) by that lake. Note len(keys) is the number of individually mapped lakes. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> from landlab.utils import StablePriorityQueue >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> open = StablePriorityQueue() >>> edges = np.array([11, 17, 23]) >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> out = lmb._fill_to_flat_with_tracking( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed ... ) >>> out == {8: deque([7]), 16: deque([15, 9, 14, 22])} True """ lakemappings = {} outlet_ID = self._grid.BAD_INDEX while True: try: c = heapq.heappop(pitq) except IndexError: try: c = openq.pop_task() outlet_ID = c except KeyError: break else: try: lakemappings[outlet_ID].append(c) # add this node to lake except KeyError: # this is the first node of a new lake lakemappings[outlet_ID] = deque([c]) cneighbors = all_neighbors[c] openneighbors = cneighbors[ np.logical_not(closedq[cneighbors]) ] # for efficiency closedq[openneighbors] = True for n in openneighbors: if fill_surface[n] <= fill_surface[c]: fill_surface[n] = fill_surface[c] heapq.heappush(pitq, n) else: openq.add_task(n, priority=fill_surface[n]) # print(np.sort(openq.tasks_currently_in_queue()), pitq) return lakemappings def _fill_to_slant_with_optional_tracking( self, fill_surface, all_neighbors, pitq, openq, closedq, ignore_overfill, track_lakes, ): """Implements the Barnes et al. algorithm to obtain a naturally draining surface over the grid. Assumes the _open and _closed lists have already been updated per Barnes algos 2&3, lns 1-7. This version runs a little more slowly to enable tracking of which nodes are linked to which outlets. Parameters ---------- fill_surface : 1-D array of length nnodes The surface to fill in LL node order. Modified in place. all_neighbors : (nnodes, max_nneighbours) array Adjacent nodes at each node. pitq : heap queue (i.e., a structured list) Current nodes known to be in a lake, if already identified. openq : StablePriorityQueue object Ordered queue of nodes remaining to be checked out by the algorithm that are known not to be in a lake. closedq : 1-D boolean array of length nnodes Nodes already or not to be explored by the algorithm. ignore_overfill : bool If False, method will raise a ValueError if adding an increment to the node's elevation would fundamentally alter the resulting drainage pattern (e.g., it would create a new outlet somewhere). If True, the elevation of the node will be changed regardless. track_lakes : bool If True, returns a dict with data on the lakes created. If false, returns an empty dict. Returns ------- lakemappings : dict If track_lakes, {outlet_ID : [nodes draining to outlet]}. This is a dict with outlet nodes of individual lakes as keys, and lists (strictly, deques) of each node inundated (i.e., depth > 0.) by that lake. Note len(keys) is the number of individually mapped lakes. If not track_lakes, an empty dict. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> from landlab.utils import StablePriorityQueue >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z.reshape(mg.shape)[2, 1:-1] = [2.0, 1.0, 0.5, 1.5] >>> z.reshape(mg.shape)[1, 1:-1] = [2.1, 1.1, 0.6, 1.6] >>> z.reshape(mg.shape)[3, 1:-1] = [2.2, 1.2, 0.7, 1.7] >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> open = StablePriorityQueue() >>> edges = np.array([11, 17, 23]) >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> out = lmb._fill_to_slant_with_optional_tracking( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False, True ... ) >>> out == {16: deque([15, 9, 8, 14, 20, 21])} True Test two pits: >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> open = StablePriorityQueue() >>> edges = np.array([11, 17, 23]) >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> out = lmb._fill_to_slant_with_optional_tracking( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False, True ... ) >>> out == {8: deque([7]), 16: deque([15, 9, 14, 22])} True >>> fr = FlowAccumulator(mg, flow_director="D4") >>> fr.run_one_step() >>> np.all(mg.at_node["flow__sink_flag"][mg.core_nodes] == 0) True >>> drainage_area = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 1.0, 2.0, 3.0, 1.0, 1.0], ... [0.0, 1.0, 4.0, 9.0, 11.0, 11.0], ... [0.0, 1.0, 2.0, 1.0, 1.0, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> np.allclose(mg.at_node["drainage_area"].reshape(mg.shape), drainage_area) True With track_lakes == False, fill still works just fine, but the dict returned is empty: >>> z[:] = mg.node_x.max() - mg.node_x # all this as above >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> open = StablePriorityQueue() >>> edges = np.array([11, 17, 23]) >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> lmb._fill_to_slant_with_optional_tracking( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False, False ... ) # empty dict now {} >>> fr.run_one_step() # drains fine still, as above >>> np.allclose(mg.at_node["drainage_area"].reshape(mg.shape), drainage_area) True Test a failing example: >>> mg = RasterModelGrid((3, 7)) >>> for edge in ("top", "right", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z.reshape(mg.shape)[1, 1:-1] = [1.0, 0.2, 0.1, 1.0000000000000004, 1.5] >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="Steepest") >>> lmb._closed = mg.zeros("node", dtype=bool) >>> lmb._closed[mg.status_at_node == mg.BC_NODE_IS_CLOSED] = True >>> open = StablePriorityQueue() >>> edges = np.array([7]) >>> for edgenode in edges: ... open.add_task(edgenode, priority=z[edgenode]) ... >>> lmb._closed[edges] = True >>> lmb._fill_to_slant_with_optional_tracking( ... z, mg.adjacent_nodes_at_node, lmb._pit, open, lmb._closed, False, True ... ) Traceback (most recent call last): ... ValueError: Pit is overfilled due to creation of two outlets as the minimum gradient gets applied. Suppress this Error with the ignore_overfill flag at component instantiation. ValueError was raised because Pit is overfilled due to creation of two outlets as the minimum gradient gets applied. Suppress this Error with the ignore_overfill flag at component instantiation. """ lakemappings = {} outlet_ID = self._grid.BAD_INDEX while True: try: topopen = openq.peek_at_task() except KeyError: noopen = True topopen = None else: noopen = False try: toppit = pitq[0] except IndexError: nopit = True toppit = None else: nopit = False # as above, DEJH is unclear how this clause triggers, so # retained but untested -> if (not (nopit or noopen)) and (topopen == toppit): # intentionally tight comparison c = openq.pop_task() outlet_ID = c self._PitTop = LARGE_ELEV elif not nopit: c = heapq.heappop(pitq) if np.isclose(self._PitTop, LARGE_ELEV): self._PitTop = fill_surface[c] if track_lakes: try: lakemappings[outlet_ID].append(c) # ^add this node to lake except KeyError: # ^this is the first node of a new lake lakemappings[outlet_ID] = deque([c]) else: try: c = openq.pop_task() # ^again, returns KeyError if empty except KeyError: break outlet_ID = c self._PitTop = LARGE_ELEV for n in all_neighbors[c]: if closedq[n]: continue else: closedq[n] = True nextval = np.nextafter(fill_surface[c], LARGE_ELEV) # as in non-tracker, DEJH believes this is redundant in LL # if self._gridclosednodes[n]: # heapq.heappush(pitq, n) if fill_surface[n] <= nextval: # formerly elif if self._PitTop < fill_surface[n] and nextval >= fill_surface[n]: if ignore_overfill: self._overfill_flag = True else: raise ValueError( "Pit is overfilled due to creation of two" " outlets as the minimum gradient gets" " applied. Suppress this Error with the" " ignore_overfill flag at component" " instantiation." ) fill_surface[n] = nextval heapq.heappush(pitq, n) else: openq.add_task(n, priority=fill_surface[n]) return lakemappings def _track_original_surface(self): """This helper method ensures that if flow is to be redircted, the _redirect_flowdirs() method can still get access to this information when it needs it. The idea here is that the operation is essentially free when surface and fill_surface were different to start with, which should make us faster. Examples -------- >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6), xy_spacing=2.0) >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z_new = mg.add_zeros("topographic__fill", at="node", dtype=float) >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="D8", ... surface="topographic__elevation", ... fill_surface="topographic__fill", ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> orig_surf = lmb._track_original_surface() >>> z is orig_surf True >>> lmb = LakeMapperBarnes( ... mg, ... method="D8", ... surface="topographic__elevation", ... fill_surface="topographic__elevation", ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> orig_surf = lmb._track_original_surface() >>> z is orig_surf False """ if self._inplace: orig_surf = self._surface.copy() else: orig_surf = self._surface return orig_surf def _redirect_flowdirs(self, surface, lake_dict, openq): """For nodes within lakes that have already been defined, modifies existing FlowDirector fields to account for the lake filling, viz. 'flow__receiver_node', 'flow__link_to_receiver_node', 'flow__sink_flag', and 'topographic__steepest_slope'. Note that the topographic__steepest_slope of a lake node will always be exactly 0., even if fill_flat is False. This is because we are adding an increment to elevation at machine precision. Examples -------- >>> import numpy as np >>> from collections import deque >>> from landlab import NodeStatus, RasterModelGrid >>> from landlab.components import ( ... LakeMapperBarnes, ... FlowDirectorSteepest, ... FlowAccumulator, ... ) >>> from landlab.utils import StablePriorityQueue >>> mg = RasterModelGrid((5, 6), xy_spacing=2.0) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[23] = 1.3 >>> z[15] = -2.0 # this deep pit causes the outlet to first drain *in* >>> z[10] = 1.3 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own, if D8 >>> z[9] = -1.0 >>> z[14] = 0.6 # [9, 14, 15] is a lake in both methods >>> z[16] = 1.2 >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 if D8 >>> z_init = z.copy() >>> fd = FlowDirectorSteepest(mg) >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=True, ... redirect_flow_steepest_descent=True, ... track_lakes=True, ... ) In this test, we won't run the lmb. Instead, directly specify the correct answer: >>> lake_dict = {8: deque([7]), 16: deque([15, 9, 14, 22])} >>> fd.run_one_step() # fill the director fields >>> fa.run_one_step() # get a drainage_area >>> np.all(mg.at_node["flow__sink_flag"][[7, 15, 22]]) # sinks True >>> nodes_in_lakes = np.array([7, 8, 9, 14, 15, 16, 22]) >>> nodes_not_in_lakes = np.setdiff1d(mg.nodes.flat, nodes_in_lakes) >>> openq = StablePriorityQueue() # empty dummy Note we're here defining the outlets as inside the lakes, which isn't actually the behaviour of the component, but helps us demonstrate what changes, below. Now save the info we already have on the Flow fields: >>> receivers_init = mg.at_node["flow__receiver_node"].copy() >>> rec_links_init = mg.at_node["flow__link_to_receiver_node"].copy() >>> steepest_init = mg.at_node["topographic__steepest_slope"].copy() >>> drainage_area = mg.at_node["drainage_area"].copy() >>> orig_surf = lmb._track_original_surface() Note flow doesn't make it to the outlets: >>> outlets = np.where(mg.status_at_node == NodeStatus.FIXED_VALUE) >>> drainage_area[outlets].sum() == mg.cell_area_at_node[mg.core_nodes].sum() False Now, run the method: >>> lmb._redirect_flowdirs(orig_surf, lake_dict, openq) Now the flow directions all ignore the pits: >>> np.all( ... mg.at_node["flow__receiver_node"].reshape(mg.shape) ... == [ ... [0, 1, 2, 3, 4, 5], ... [6, 8, 9, 15, 9, 11], ... [12, 14, 15, 16, 17, 17], ... [18, 20, 14, 15, 16, 23], ... [24, 25, 26, 27, 28, 29], ... ] ... ) True (Note the filling of the pits might redirect the occasional node not in a lake, but on its perimeter - if the node used to drain into the lake, but now has a steeper descent path elsewhere.) There are now no pits: >>> np.any(mg.at_node["flow__sink_flag"][[7, 15, 22]]) False The lake nodes now flow out: >>> mg.at_node["flow__receiver_node"][lake_dict[16]] array([16, 15, 15, 16]) >>> mg.at_node["flow__receiver_node"][lake_dict[8]] array([8]) ...and any outlet nodes that used to drain into the lake now drain out. >>> receivers_init[16] 15 >>> mg.at_node["flow__receiver_node"][16] 17 >>> np.isclose(mg.at_node["topographic__steepest_slope"][16], 0.6) True If we reaccumulate the flow, we'll now see that the boundary nodes do now accumulate the total available discharge: >>> area, discharge = fa.accumulate_flow(update_flow_director=False) >>> mg.at_node["drainage_area"][outlets].sum() == ( ... mg.cell_area_at_node[mg.core_nodes].sum() ... ) True """ closedq = self._grid.ones(at="node", dtype=int) # Using a slightly different approach. We recognise three types: lake # (0), lake margin (1), and closed (2). This lets us work the # perimeter too. Open each lake as needed. # close the known boundary nodes: closedq[self._grid.status_at_node != NodeStatus.CORE] = 2 # now the actual loop. Work forward lake by lake to avoid unnecessary # processing (nodes outside lakes are already correct, by definition). for outlet, lakenodes in lake_dict.items(): # open the lake: closedq[lakenodes] = 0 # make a deque for liminal nodes: liminal_nodes = deque([]) openq.add_task(outlet, priority=surface[outlet]) # it's possible the outlet used to drain *into* the lake, # so it needs separate consideration. Likewise, the gradients # of the perimeter nodes are likely to be wrong. if self._grid.status_at_node[outlet] != NodeStatus.CORE: # don't do anything if the outlet happens to be a boundary pass else: out_elev = LARGE_ELEV for neighbor_set, link_set in zip( self._neighbor_arrays, self._link_arrays ): not_lake_neighbors = np.not_equal(closedq[neighbor_set[outlet]], 0) minusones = np.equal(neighbor_set[outlet], -1) not_lake_neighbors[minusones] = False closednodes = np.equal( self._grid.status_at_node[neighbor_set[outlet]], self._grid.BC_NODE_IS_CLOSED, ) # closed BCs can't count not_lake_neighbors[closednodes] = False try: min_val = np.amin( surface[neighbor_set[outlet][not_lake_neighbors]] ) except ValueError: continue if min_val < out_elev: viable_nodes = neighbor_set[outlet][not_lake_neighbors] min_neighbor_byTrue = np.argmin(surface[viable_nodes]) min_neighbor = viable_nodes[min_neighbor_byTrue] min_link = link_set[outlet][not_lake_neighbors][ min_neighbor_byTrue ] out_elev = min_val self._receivers[outlet] = min_neighbor self._receiverlinks[outlet] = min_link self._steepestslopes[outlet] = ( surface[outlet] - surface[min_neighbor] ) / self._neighbor_lengths[min_link] while True: try: c = openq.pop_task() except KeyError: break else: closedq[c] = 2 # close it # if raster, do the neighbors & diagonals separate... for neighbor_set, link_set in zip( self._neighbor_arrays, self._link_arrays ): for n, l in zip(neighbor_set[c, :], link_set[c, :]): # fully closed if (closedq[n] == 2) or (n == -1): continue elif self._grid.status_at_node[n] != NodeStatus.CORE: closedq[n] = 2 continue else: if closedq[n] == 0: self._receivers[n] = c self._receiverlinks[n] = l self._steepestslopes[n] = 0.0 closedq[n] = 2 # close it openq.add_task(n, priority=surface[n]) else: # it's liminal (1); grads likely wrong # ...but it's not if set by the outlet... if c == outlet: # still need these nodes to be explored # by other lake nodes as needed, so # don't close either pass else: # liminal to actual lake closedq[n] = 2 liminal_nodes.append(n) # ...& don't add to the queue # TODO: obvious case for Cython accel here # Now know which nodes we need to reassess. So: for liminal in liminal_nodes: min_elev = LARGE_ELEV min_link = -1 for neighbor_set, link_set in zip( self._neighbor_arrays, self._link_arrays ): neighbors = neighbor_set[liminal] neighbors_valid = np.not_equal(neighbors, -1) closednodes = np.equal( self._grid.status_at_node[neighbors], self._grid.BC_NODE_IS_CLOSED, ) # closed BCs can't count neighbors_valid[closednodes] = False neighbors_to_check = neighbors[neighbors_valid] if len(neighbors_to_check) == 0: continue else: min_neighbor_now = np.amin( self._fill_surface[neighbors_to_check] ) if min_neighbor_now < min_elev: min_elev = min_neighbor_now links_available = link_set[liminal][neighbors_valid] min_link_of_valid = np.argmin( self._fill_surface[neighbors_to_check] ) min_receiver = neighbors_to_check[min_link_of_valid] min_link = links_available[min_link_of_valid] max_grad = ( self._fill_surface[liminal] - min_elev ) / self._neighbor_lengths[min_link] else: pass assert min_link != -1, neighbors_valid # ^link successfully found self._receivers[liminal] = min_receiver self._receiverlinks[liminal] = min_link self._steepestslopes[liminal] = max_grad # by the time we get here, we've removed all the pits! So... self._grid.at_node["flow__sink_flag"][lakenodes] = 0 # reclose the lake: closedq[outlet] = 1 closedq[lakenodes] = 1 closedq[liminal_nodes] = 1
[docs] def update(self): """Alias for running one step.""" self.run_one_step()
[docs] def run_one_step(self): """Fills the surface to fill all pits. Note that a second run on a surface that has already been filled will *not* "see" any existing. lakes correctly - it will see lakes, but with zero depths. In particular, if fill_flat is False, an attempt to fill a surface a second time will raise a ValueError unless ignore_overfill. (In this case, setting ignore_overfill is True will give the expected behaviour.) If reaccumulate_flow was True at instantiation, run_one_step also updates all the various flow fields produced by the FlowDirector and FlowAccumulator components. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> from landlab.components import FlowDirectorSteepest >>> mg = RasterModelGrid((5, 6), xy_spacing=2.0) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> mg.at_node["topographic__elevation"] = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 2.1, 1.1, 0.6, 1.6, 0.0], ... [0.0, 2.0, 1.0, 0.5, 1.5, 0.0], ... [0.0, 2.2, 1.2, 0.7, 1.7, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> z = mg.at_node["topographic__elevation"] >>> z_init = mg.at_node["topographic__elevation"].copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) TODO: once return_array_at_node is fixed, this example should also take fill_surface... >>> lmb.run_one_step() >>> z_out = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 2.1, 1.5, 1.5, 1.6, 0.0], ... [0.0, 2.0, 1.5, 1.5, 1.5, 0.0], ... [0.0, 2.2, 1.5, 1.5, 1.7, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> np.allclose(z.reshape(mg.shape), z_out) True >>> try: ... lmb.lake_map # not created, as we aren't tracking ... except ValueError: ... print( ... "ValueError was raised:" ... " Enable tracking to access information about lakes" ... ) ... ValueError was raised: Enable tracking to access information about lakes >>> lmb.was_there_overfill # everything fine with slope adding False >>> fd = FlowDirectorSteepest(mg) >>> fa = FlowAccumulator(mg) # routing will work fine now >>> fd.run_one_step() >>> fa.run_one_step() >>> np.all(mg.at_node["flow__sink_flag"][mg.core_nodes] == 0) True >>> drainage_area = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 4.0, 8.0, 12.0, 4.0, 4.0], ... [0.0, 4.0, 16.0, 36.0, 40.0, 40.0], ... [0.0, 4.0, 8.0, 4.0, 4.0, 4.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> np.allclose(mg.at_node["drainage_area"].reshape(mg.shape), drainage_area) True Test two pits: >>> z[:] = mg.node_x.max() - mg.node_x >>> z[23] = 1.3 >>> z[15] = 0.3 >>> z[10] = 1.3 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own, if D8 >>> z[9] = 0.5 >>> z[14] = 0.6 # [9, 14, 15] is a lake in both methods >>> z[16] = 1.2 >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 if D8 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes(mg, method="D8", fill_flat=True, track_lakes=True) >>> lmb.run_one_step() # note the D8 routing now >>> lmb.lake_dict == {22: deque([15, 9, 14])} True >>> lmb.number_of_lakes 1 >>> lmb.lake_depths # z was both surface and "fill_surface" Traceback (most recent call last): ... ValueError: surface and fill_surface must be different fields or arrays to enable the property lake_depths! ValueError was raised because surface and fill_surface must be different fields or arrays to enable the property fill_depth! >>> z[:] = z_init >>> lmb = LakeMapperBarnes( ... mg, method="Steepest", fill_flat=False, track_lakes=True ... ) >>> lmb.run_one_step() # compare to the method='D8' lakes, above... >>> lmb.lake_dict == {8: deque([7]), 16: deque([15, 9, 14, 22])} True >>> lmb.number_of_lakes 2 >>> np.allclose(lmb.lake_areas, np.array([16.0, 4.0])) True >>> lmb.run_one_step() Traceback (most recent call last): ... ValueError: Pit is overfilled due to creation of two outlets as the minimum gradient gets applied. Suppress this Error with the ignore_overfill flag at component instantiation. ValueError was raised because Pit is overfilled due to creation of two outlets as the minimum gradient gets applied. Suppress this Error with the ignore_overfill flag at component instantiation. Suppress this behavior with ignore_overfill: >>> z[:] = z_init >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... track_lakes=True, ... ignore_overfill=True, ... ) >>> lmb.run_one_step() >>> lmb.lake_dict == {8: deque([7]), 16: deque([15, 9, 14, 22])} True >>> lmb.run_one_step() >>> np.allclose(lmb.lake_areas, np.array([16.0, 4.0])) # found them! True The component can redirect flow to account for the fills that have been carried out (all necessary fields get updated): >>> z[:] = z_init >>> fd.run_one_step() >>> init_flowdirs = mg.at_node["flow__receiver_node"].copy() >>> fa.run_one_step() >>> init_areas = mg.at_node["drainage_area"].copy() >>> init_qw = mg.at_node["surface_water__discharge"].copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... track_lakes=True, ... redirect_flow_steepest_descent=False, ... ignore_overfill=True, ... ) >>> lmb.run_one_step() >>> np.all(mg.at_node["flow__receiver_node"] == init_flowdirs) True >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... track_lakes=True, ... redirect_flow_steepest_descent=True, ... ignore_overfill=True, ... ) >>> lmb.run_one_step() >>> np.all(mg.at_node["flow__receiver_node"] == init_flowdirs) False However, note that unless the reaccumulate_flow argument is also set, the 'drainage_area' and 'surface_water__discharge' fields *won't* also get updated: >>> np.all(mg.at_node["drainage_area"] == init_areas) True >>> np.all(mg.at_node["surface_water__discharge"] == init_qw) True >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... track_lakes=True, ... redirect_flow_steepest_descent=True, ... reaccumulate_flow=True, ... ignore_overfill=True, ... ) >>> lmb.run_one_step() >>> np.all(mg.at_node["drainage_area"] == init_areas) False >>> np.all(mg.at_node["surface_water__discharge"] == init_qw) False Be sure to set both redirect_flow_steepest_descent and reaccumulate_flow to True if you want to reaccumulate flow... >>> try: ... lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... track_lakes=True, ... redirect_flow_steepest_descent=False, ... reaccumulate_flow=True, ... ignore_overfill=True, ... ) ... except ValueError: ... print("Oops!") ... Oops! The component is completely happy with irregular grids: >>> from landlab import HexModelGrid, FieldError >>> hmg = HexModelGrid((5, 4), spacing=2.0) >>> z_hex = hmg.add_zeros("topographic__elevation", at="node") >>> z_hex[:] = hmg.node_x >>> z_hex[11] = -3.0 >>> z_hex[12] = -1.0 >>> z_hex_init = z_hex.copy() >>> z_hex array([ 2., 4., 6., 8., 1., 3., 5., 7., 9., 0., 2., -3., -1., 8., 10., 1., 3., 5., 7., 9., 2., 4., 6., 8.]) As you can see, nodes 11 and 12 are now a pit. If they were to fill they would fill to the level of 2, the lowest downstream value. >>> fa = FlowAccumulator(hmg) >>> lmb = LakeMapperBarnes( ... hmg, method="Steepest", fill_flat=True, track_lakes=False ... ) >>> lmb.run_one_step() >>> np.allclose(z_hex[10:13], 2.0) True >>> hmg = HexModelGrid((5, 4), spacing=2.0) >>> z_hex = hmg.add_zeros("topographic__elevation", at="node") >>> z_hex[:] = z_hex_init >>> try: ... lmb = LakeMapperBarnes( ... hmg, ... method="Steepest", ... fill_flat=False, ... surface=z_hex_init, ... redirect_flow_steepest_descent=True, ... track_lakes=True, ... ) ... except FieldError: ... print("Oops!") # flowdir field must already exist! ... Oops! >>> fd = FlowDirectorSteepest(hmg) >>> fa = FlowAccumulator(hmg) >>> lmb = LakeMapperBarnes( ... hmg, ... method="Steepest", ... fill_flat=False, ... surface=z_hex_init, ... redirect_flow_steepest_descent=True, ... track_lakes=True, ... ) >>> fd.run_one_step() >>> lmb.run_one_step() >>> np.allclose(z_hex[10:13], 2.0) True >>> z_hex[11] > z_hex[10] True >>> z_hex[12] > z_hex[11] True >>> np.allclose(lmb.lake_depths[10:14], np.array([0.0, 5.0, 3.0, 0.0])) True >>> np.testing.assert_array_almost_equal(lmb.lake_volumes, 27.712, decimal=3) Together, all this means that we can now run a topographic growth model that permits flooding as it runs: >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> from landlab.components import FlowDirectorSteepest >>> from landlab.components import FastscapeEroder >>> mg = RasterModelGrid((6, 8)) >>> for edge in ("right", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... Because it is what we want the FastscapeEroder to see and work on, it's actually the water surface that needs to go in as 'topographic__elevation'. We'll also need to keep track of the bed elevation though, since the LakeMapper will need it. We start them equal (i.e., topo starts dry). >>> z_water = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z_water[:] = mg.node_x >>> z_water[11] = 1.5 >>> z_water[19] = 0.5 >>> z_water[34] = 1.1 >>> z_bed = mg.add_zeros("bedrock__elevation", at="node", dtype=float) >>> z_bed[:] = z_water # topo starts dry Let's just take a look: >>> np.all( ... np.equal( ... np.round(z_water, 2).reshape(mg.shape), ... [ ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 1.5, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 0.5, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 1.1, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... ], ... ) ... ) True >>> fd = FlowDirectorSteepest(mg) >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="D8", ... fill_flat=True, ... surface="bedrock__elevation", ... fill_surface="topographic__elevation", ... redirect_flow_steepest_descent=True, ... reaccumulate_flow=True, ... track_lakes=True, ... ) >>> sp = FastscapeEroder(mg, K_sp=1.0, m_sp=0.0, n_sp=1.0) >>> fd.run_one_step() >>> fa.run_one_step() # node 18 is draining into the pit... >>> np.isclose(mg.at_node["topographic__steepest_slope"][18], 1.5) True >>> np.allclose( ... mg.at_node["drainage_area"].reshape(mg.shape), ... [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [2.0, 2.0, 1.0, 4.0, 3.0, 2.0, 1.0, 0.0], ... [1.0, 1.0, 1.0, 13.0, 3.0, 2.0, 1.0, 0.0], ... [2.0, 2.0, 1.0, 4.0, 3.0, 2.0, 1.0, 0.0], ... [6.0, 6.0, 5.0, 4.0, 3.0, 2.0, 1.0, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ], ... ) True >>> lmb.run_one_step() # now node 18 drains correctly, outward -> >>> np.isclose(mg.at_node["topographic__steepest_slope"][18], 1.0) True >>> np.allclose( ... mg.at_node["drainage_area"].reshape(mg.shape), ... [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [13.0, 13.0, 12.0, 4.0, 3.0, 2.0, 1.0, 0.0], ... [2.0, 2.0, 1.0, 7.0, 3.0, 2.0, 1.0, 0.0], ... [2.0, 2.0, 1.0, 1.0, 3.0, 2.0, 1.0, 0.0], ... [7.0, 7.0, 6.0, 4.0, 3.0, 2.0, 1.0, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ], ... ) True >>> np.all( ... np.equal( ... np.round(z_water, 2).reshape(mg.shape), ... [ ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 2.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 2.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 1.1, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... ], ... ) ... ) True >>> sp.run_one_step(0.05) # note m=0 to illustrate effect of slopes >>> np.all( ... np.equal( ... np.round(z_water, 2).reshape(mg.shape), ... [ ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 0.95, 1.95, 2.0, 3.9, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.95, 2.0, 3.9, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.95, 2.93, 3.93, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.09, 2.91, 3.95, 4.95, 5.95, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... ], ... ) ... ) True If we want to keep this going honouring the depths of the lakes try this next in your loop: >>> z_bed[:] = np.minimum(z_water, z_bed) >>> np.all( ... np.equal( ... np.round(z_bed, 2).reshape(mg.shape), ... [ ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 0.95, 1.95, 1.5, 3.9, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.95, 0.5, 3.9, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.95, 2.93, 3.93, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.09, 2.91, 3.95, 4.95, 5.95, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... ], ... ) ... ) True >>> fd.run_one_step() >>> fa.run_one_step() >>> lmb.run_one_step() Lake node depths are now updated in lmb: >>> np.round([lmb.lake_depths[lake] for lake in lmb.lake_dict.values()], 2) array([[0.45, 1.45]]) ...and the "topography" (i.e., water surface) at the flooded nodes has lowered itself as the lip of the outlet was eroded in the last step: >>> np.all( ... np.equal( ... np.round(z_water, 2).reshape(mg.shape), ... [ ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... [0.0, 0.95, 1.95, 1.95, 3.9, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.95, 1.95, 3.9, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.95, 2.93, 3.93, 4.95, 5.95, 7.0], ... [0.0, 0.95, 1.09, 2.91, 3.95, 4.95, 5.95, 7.0], ... [0.0, 1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0], ... ], ... ) ... ) True >>> sp.run_one_step(0.05) ...and so on. Note that this approach, without passing `flooded_nodes` to the FastscapeEroder run method, is both more "Landlabbic" and also ensures the information about the lake and the water surface topography are all updated cleanly and correctly. """ if "flow__receiver_node" in self._grid.at_node and self._grid.at_node[ "flow__receiver_node" ].size != self._grid.size("node"): raise NotImplementedError( "A route-to-multiple flow director has been " "run on this grid. The landlab development team has not " "verified that LakeMapperBarnes is compatible with " "route-to-multiple methods. Please open a GitHub Issue " "to start this process." ) # do the prep: # create the StasblePriorityQueue locaslly to permit garbage collection _open = StablePriorityQueue() # increment the run counter self._runcount = next(self._runcounter) # First get _fill_surface in order. self._fill_surface[:] = self._surface # surfaces begin identical # note this is nice & efficent if _fill_surface is _surface # if we're doing a redirect, we're going to need to preserve this # initial topo, so let's do that: if not self._dontredirect: orig_topo = self._track_original_surface() # now, return _closed to its initial cond, w only the BC_NODE_IS_CLOSED # and grid draining nodes pre-closed: closedq = self._closed.copy() if self._track_lakes: for edgenode in self._edges: _open.add_task(edgenode, priority=self._surface[edgenode]) closedq[self._edges] = True if self._fill_flat: self._lakemappings = self._fill_to_flat_with_tracking( self._fill_surface, self._allneighbors, self._pit, _open, closedq, ) else: self._lakemappings = self._fill_to_slant_with_optional_tracking( self._fill_surface, self._allneighbors, self._pit, _open, closedq, ignore_overfill=self._ignore_overfill, track_lakes=True, ) if not self._dontredirect: self._redirect_flowdirs(orig_topo, self._lakemappings, _open) if self._reaccumulate: _, _ = self._fa.accumulate_flow(update_flow_director=False) else: # not tracked # note we've already checked _dontredirect is True in setup, # so we don't need to worry about these cases. for edgenode in self._edges: _open.add_task(edgenode, priority=self._surface[edgenode]) closedq[self._edges] = True while True: try: self._fill_one_node( self._fill_surface, self._allneighbors, self._pit, _open, closedq, self._ignore_overfill, ) except KeyError: # run out of nodes to fill... break
@property def lake_dict(self): """Return a dictionary where the keys are the outlet nodes of each lake, and the values are deques of nodes within each lake. Items are not returned in ID order. The outlet nodes are NOT part of the lake. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> # z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> # z.reshape(mg.shape)[2, 1:-1] = [2.0, 1.0, 0.5, 1.5] >>> # z.reshape(mg.shape)[1, 1:-1] = [2.1, 1.1, 0.6, 1.6] >>> # z.reshape(mg.shape)[3, 1:-1] = [2.2, 1.2, 0.7, 1.7] >>> mg.at_node["topographic__elevation"] = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 2.1, 1.1, 0.6, 1.6, 0.0], ... [0.0, 2.0, 1.0, 0.5, 1.5, 0.0], ... [0.0, 2.2, 1.2, 0.7, 1.7, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> z = mg.at_node["topographic__elevation"] >>> z_init = mg.at_node["topographic__elevation"].copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_surface=z, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> lmb.run_one_step() >>> try: ... lmb.lake_dict ... except ValueError: ... print( ... "ValueError was raised: " ... + "Enable tracking to access information about lakes" ... ) ... ValueError was raised: Enable tracking to access information about lakes >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_surface=z, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.lake_dict == {16: deque([15, 9, 8, 14, 20, 21])} True """ if not self._track_lakes: raise ValueError("Enable tracking to access information about lakes") return self._lakemappings @property def lake_outlets(self): """Returns the outlet for each lake, not necessarily in ID order. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> mg.at_node["topographic__elevation"] = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 2.1, 1.1, 0.6, 1.6, 0.0], ... [0.0, 2.0, 1.0, 0.5, 1.5, 0.0], ... [0.0, 2.2, 1.2, 0.7, 1.7, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> z = mg.at_node["topographic__elevation"] >>> z_init = mg.at_node["topographic__elevation"].copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_surface=z, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> lmb.run_one_step() >>> try: ... lmb.lake_outlets ... except ValueError: ... print( ... "ValueError was raised:" ... " Enable tracking to access information about lakes" ... ) ... ValueError was raised: Enable tracking to access information about lakes >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_surface=z, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.lake_outlets == [16] True """ if not self._track_lakes: raise ValueError("Enable tracking to access information about lakes") return list(self._lakemappings.keys()) @property def number_of_lakes(self): """Return the number of individual lakes. Lakes sharing outlet nodes are considered part of the same lake. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_surface=z, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> lmb.run_one_step() >>> try: ... lmb.number_of_lakes ... except ValueError: ... print( ... "ValueError was raised:" ... " Enable tracking to access information about lakes" ... ) ... ValueError was raised: Enable tracking to access information about lakes >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... surface=z_init, ... fill_surface=z, ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.number_of_lakes 2 """ if not self._track_lakes: raise ValueError("Enable tracking to access information about lakes") return len(self._lakemappings) @property def lake_map(self): """Return an array of ints, where each node within a lake is labelled with its outlet node ID. The outlet nodes are NOT part of the lakes. Nodes not in a lake are labelled with BAD_INDEX_VALUE (default -1). Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> lmb.run_one_step() >>> try: ... lmb.lake_map ... except ValueError: ... print( ... "ValueError was raised:" ... " Enable tracking to access information about lakes" ... ) ... ValueError was raised: Enable tracking to access information about lakes >>> z[:] = z_init >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lake_map = [ ... [-1, -1, -1, -1, -1, -1], ... [-1, 8, -1, 16, -1, -1], ... [-1, -1, 16, 16, -1, -1], ... [-1, -1, -1, -1, 16, -1], ... [-1, -1, -1, -1, -1, -1], ... ] >>> np.all(lmb.lake_map.reshape(mg.shape) == lake_map) True Note that the outlet node is NOT part of the lake. Updating the elevations works fine: >>> z.reshape(mg.shape)[1:4, 1:-1] = [ ... [2.1, 1.1, 0.6, 1.6], ... [2.0, 1.0, 0.5, 1.5], ... [2.2, 1.2, 0.7, 1.7], ... ] >>> lmb.run_one_step() >>> new_lake_map = [ ... [-1, -1, -1, -1, -1, -1], ... [-1, -1, 16, 16, -1, -1], ... [-1, -1, 16, 16, -1, -1], ... [-1, -1, 16, 16, -1, -1], ... [-1, -1, -1, -1, -1, -1], ... ] >>> np.all(lmb.lake_map.reshape(mg.shape) == new_lake_map) True """ if self._runcount > self._lastcountforlakemap: # things have changed since last call to lake_map self._lake_map = np.full( self._grid.number_of_nodes, self._grid.BAD_INDEX, dtype=int ) for outlet, lakenodes in self.lake_dict.items(): self._lake_map[lakenodes] = outlet else: pass # old map is fine self._lastcountforlakemap = self._runcount return self._lake_map @property def lake_at_node(self): """Return a boolean array, True if the node is flooded, False otherwise. The outlet nodes are NOT part of the lakes. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lake_at_node = np.array( ... [ ... [0, 0, 0, 0, 0, 0], ... [0, 1, 0, 1, 0, 0], ... [0, 0, 1, 1, 0, 0], ... [0, 0, 0, 0, 1, 0], ... [0, 0, 0, 0, 0, 0], ... ], ... dtype=bool, ... ) >>> np.all(lmb.lake_at_node.reshape(mg.shape) == lake_at_node) True """ return self.lake_map != self._grid.BAD_INDEX @property def lake_depths(self): """Return the change in surface elevation at each node this step. Requires that surface and fill_surface were not the same array at instantiation. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() This won't work as surface & fill_surface are both z >>> lmb.lake_depths Traceback (most recent call last): ... ValueError: surface and fill_surface must be different fields or arrays to enable the property lake_depths! `ValueError` was raised because surface and fill_surface must be different fields or arrays to enable the property lake_depths! >>> z[:] = z_init >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... surface=z_init, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lake_depths = [ ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... [0.0, 1.0, 0.0, 0.5, 0.0, 0.0], ... [0.0, 0.0, 0.4, 0.7, 0.0, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.1, 0.0], ... [0.0, 0.0, 0.0, 0.0, 0.0, 0.0], ... ] >>> np.all( ... np.equal(lmb.lake_depths.reshape(mg.shape), lake_depths) ... ) # slope applied, so... False >>> np.allclose(lmb.lake_depths.reshape(mg.shape), lake_depths) True """ if self._inplace: raise ValueError( "surface and fill_surface must be different fields or " + "arrays to enable the property lake_depths!" ) return self._fill_surface - self._surface @property def lake_areas(self): """A nlakes-long array of the area of each lake. The order is the same as that of the keys in lake_dict, and of lake_outlets. Note that outlet nodes are not parts of the lakes. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=False, ... ) >>> lmb.run_one_step() >>> try: ... lmb.lake_areas # note track_lakes=False ... except ValueError: ... print( ... "ValueError was raised: " ... + "Enable tracking to access information about lakes" ... ) ... ValueError was raised: Enable tracking to access information about lakes >>> z[:] = z_init >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.lake_outlets [16, 8] >>> np.allclose(lmb.lake_areas, np.array([4.0, 1.0])) True """ lakeareas = np.empty(self.number_of_lakes, dtype=float) for i, lakenodes in enumerate(self.lake_dict.values()): lakeareas[i] = self._grid.cell_area_at_node[lakenodes].sum() return lakeareas @property def lake_volumes(self): """A nlakes-long array of the volume of each lake. The order is the same as that of the keys in lake_dict, and of lake_outlets. Note that this calculation is performed relative to the initial surface, so is only a true lake volume if the initial surface was the rock suface (not an earlier water level). Requires that surface and fill_surface were not the same array at instantiation. Examples -------- >>> import numpy as np >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((5, 6)) >>> for edge in ("left", "top", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z[:] = mg.node_x.max() - mg.node_x >>> z[[10, 23]] = 1.1 # raise "guard" exit nodes >>> z[7] = 2.0 # is a lake on its own >>> z[9] = 0.5 >>> z[15] = 0.3 >>> z[14] = 0.6 # [9, 14, 15] is a lake >>> z[22] = 0.9 # a non-contiguous lake node also draining to 16 >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.lake_volumes # won't work as surface & fill_surface are both z Traceback (most recent call last): ... ValueError: surface and fill_surface must be different fields or arrays to enable the property lake_depths! >>> z[:] = z_init >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... surface=z_init, ... redirect_flow_steepest_descent=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.lake_outlets [16, 8] >>> np.allclose(lmb.lake_volumes, np.array([1.7, 1.0])) True """ lake_vols = np.empty(self.number_of_lakes, dtype=float) col_vols = self._grid.cell_area_at_node * self.lake_depths for i, lakenodes in enumerate(self.lake_dict.values()): lake_vols[i] = col_vols[lakenodes].sum() return lake_vols @property def was_there_overfill(self): """If the ignore_overfill flag was set to True at instantiation, this property indicates if any depression in the grid has, at any point, been overfilled. Examples -------- >>> from landlab import RasterModelGrid >>> from landlab.components import LakeMapperBarnes, FlowAccumulator >>> mg = RasterModelGrid((3, 7)) >>> for edge in ("top", "right", "bottom"): ... mg.status_at_node[mg.nodes_at_edge(edge)] = mg.BC_NODE_IS_CLOSED ... >>> z = mg.add_zeros("topographic__elevation", at="node", dtype=float) >>> z.reshape(mg.shape)[1, 1:-1] = [1.0, 0.2, 0.1, 1.0000000000000004, 1.5] >>> z_init = z.copy() >>> fa = FlowAccumulator(mg) >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=True, ... redirect_flow_steepest_descent=False, ... ignore_overfill=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.was_there_overfill Traceback (most recent call last): ... ValueError: was_there_overfill is only defined if filling to an inclined surface! `ValueError` was raised because was_there_overfill is only defined if filling to an inclined surface! >>> z_init = z.copy() >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... ignore_overfill=False, ... track_lakes=True, ... ) >>> lmb.run_one_step() Traceback (most recent call last): ... ValueError: Pit is overfilled due to creation of two outlets as the minimum gradient gets applied. Suppress this Error with the ignore_overfill flag at component instantiation. `ValueError` was raised because Pit is overfilled due to creation of two outlets as the minimum gradient gets applied. Suppress this Error with the ignore_overfill flag at component instantiation. >>> z_init = z.copy() >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... ignore_overfill=True, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.was_there_overfill True >>> z.reshape(mg.shape)[1, 1:-1] = [1.0, 0.2, 0.1, 1.0, 1.5] >>> lmb.run_one_step() >>> lmb.was_there_overfill # still true as was in the previous example True >>> z.reshape(mg.shape)[1, 1:-1] = [1.0, 0.2, 0.1, 1.0, 1.5] >>> lmb = LakeMapperBarnes( ... mg, ... method="Steepest", ... fill_flat=False, ... redirect_flow_steepest_descent=False, ... ignore_overfill=True, ... track_lakes=True, ... ) >>> lmb.run_one_step() >>> lmb.was_there_overfill # Now reset False >>> lmb.run_one_step() # 2nd run on same fill_surface creates overfill >>> lmb.was_there_overfill True Note however that in this last example, values have NOT changed. """ if self._fill_flat is True: raise ValueError( "was_there_overfill is only defined if filling to an " + "inclined surface!" ) return self._overfill_flag